| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Reproductive Biology |
2Departments of Biology and Mathematics, Grinnell College, Grinnell, Iowa 50112 USA; 3Department of Biology, Grinnell College, Grinnell, Iowa 50112 USA
Received for publication December 17, 2002. Accepted for publication April 24, 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: Lobelia cardinalis • Lobelia siphilitica • Lobeliaceae • natural selection • phenotypic selection • pollination • spatial variation • temporal variation
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). Within a species, variation in the biotic and abiotic environment may result in spatial or temporal variation in the strength and direction of natural selection. All things being equal, these patterns of variation in selection will have strikingly different consequences for the evolution of phenotypic diversity in natural populations. In the absence of plasticity, variation in selection among populations should lead to genetically based phenotypic differences in the trait under selection. In contrast, high temporal variation in selection should prevent this genetic differentiation (e.g., Levins, 1968
; Bell, 1997
; Reboud and Bell, 1997
). As such, determining the relative importance of spatial vs. temporal variation in natural selection is essential for making testable predictions about the distribution of phenotypes across natural populations. Any difference between predicted and observed phenotypes suggests that adaptive evolution is constrained or that other mechanisms of evolution, such as gene flow or drift, are operating.
The great diversity of floral displays in nature makes these traits one of the model systems for studying adaptive evolution. Although natural selection has been measured for floral traits of many species, significant spatial variation (Gilbert et al., 1996
; O'Connell and Johnston, 1998
; Caruso, 2000
, 2001
; Totland, 2001
) and temporal variation (Schemske and Horvitz, 1989
; Campbell, 1991
; Gross et al., 1998
; Maad, 2000
) in selection has been documented in relatively few systems. The ecological causes of this variation are rarely well studied (but see Caruso, 2000
; Totland, 2001
), but generally involve indirect effects of the biotic or abiotic environment on plant–pollinator interactions. For example, Totland (2001)
found that altitudinal differences in selection on floral traits of Ranunculus acris were caused by differences in temperature, which influenced both pollinator behavior and seed set.
The goals of our study were to characterize spatial and temporal variation in selection on floral traits and to explore potential ecological causes of this variation, using the temperate wildflowers Lobelia cardinalis L. and L. siphilitica L. as a study system. Significant natural selection on floral traits has been documented in natural populations of L. cardinalis and L. siphilitica (Johnston, 1991a
). There are anecdotal reports of variation in the abiotic soil and light environment among conspecific populations of both species (Johnston, 1991a
; Pigliucci and Schlichting, 1995
; Pigliucci et al., 1997
). Johnston (1991a)
also noted variation among L. cardinalis populations in the presence of coflowering Impatiens capensis, a component of the biotic environment. We studied natural populations of L. cardinalis and L. siphilitica to answer the following questions: (1) What is the relative importance of spatial vs. temporal variation in phenotypic selection on floral traits of L. cardinalis and L. siphilitica? (2) Is spatial and temporal variation in the abiotic soil microenvironment associated with variation in selection on Lobelia?
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
and references therein). Lobelia cardinalis has 4-cm-long red flowers (Fig. 1A) pollinated by ruby-throated hummingbirds (Archilochus colubris) throughout eastern North America (Baker, 1975
; Bertin, 1982
). In contrast, L. siphilitica has 3-cm-long blue flowers (Fig. 1B) pollinated by Bombus spp. (Beaudoin Yetter, 1989
). Flowers of both species are protandrous, and pollen is shed from a tube formed by the fused anthers and filaments (Johnston, 1991a
). Both L. cardinalis and L. siphilitica are self-compatible (Johnston, 1992
), but the complete separation between staminate and pistillate phases of flower development ensures that self-fertilization is the result of geitonogamy rather than autogamy (Johnston, 1991b
). Although L. siphilitica can be gynodioecious (Beaudoin Yetter, 1989
), females were very rare in our study populations (1–9%; C. M. Caruso, unpublished data).
|
|
) and L. siphilitica (Beaudoin Yetter, 1989
) can produce clonal offshoots, but we attempted to include only one ramet per genet in the sample from each population. In addition, because both species are short-lived perennials, some genets may have been included in the sample in both 1999 and 2000. However, the location of the densest aggregation of flowering individuals within each population varied between years (C. M. Caruso, personal observation), which should minimize the possibility of resampling genets.
We measured six floral traits on each plant: corolla lobe length, corolla lobe width, corolla tube length, corolla tube width, stigma–nectary distance, and total flower number (Fig. 1). Floral morphology was measured for 10 (in 1999) or five (in 2000) flowers per plant using hand-held digital calipers. On the same day that floral morphology was measured, we estimated total flower number as the sum of the senescent flowers, open flowers, and unopened buds on each plant. An additional floral trait (stigma exsertion) was estimated by subtracting corolla length from stigma–nectary distance (Fig. 1). Stigma–nectary distance was measured because it was under significant selection in Michigan populations of Lobelia (Johnston, 1991a
). The other six floral traits were measured because they were under significant selection in other plant species (e.g., Campbell, 1991
; Conner et al., 1996
; Caruso, 2000
). Female fitness was estimated for each plant as the number of fruits per plant, which is a good predictor of total seed production in Lobelia (Johnston, 1991b
). Plants that did not produce any fruit were included in all analyses. Supplemental hand-pollination increases the number of fruits per plant in both L. cardinalis and L. siphilitica (Johnston, 1991b
), indicating that variation in fruit set among individuals could reflect variation in pollination.
Abiotic environment
We used time domain reflectometry to measure the volumetric water content (VWC) of the soil at each study population on five dates between 23 September 2000 and 25 July 2001. The VWC was estimated at 14–39 haphazardly chosen points in each population using a CS620 handheld water content sensor (Campbell Scientific, Logan, Utah, USA). Most soils are saturated at a VWC of 50% (Jackson et al., 2000
).
Data analysis
Fitness measures and floral traits
To determine if fitness measures or floral traits differed among conspecific populations measured in 2000, we used one-way analysis of variance (ANOVA) with population as the independent variable. Data from each species were analyzed using separate ANOVAs. When the ANOVA was significant, we used Tukey's HSD test to detect pairwise differences between populations. Differences in fitness measures or floral traits between (1) years in CB and (2) CERA and Krumm in 1999 were determined using a t test. The assumptions of normality and homogeneity of variance for all t tests and ANOVAs presented in this paper were tested using Lilliefors' (Wilkinson, 1997
) and Levene's (Underwood, 1997
) tests, respectively.
Directional selection
We estimated phenotypic selection for each floral trait, population, and year. Based on the guideline that sample size should be at least 10x the number of independent variables in a multiple regression (Mitchell, 2001
), the sample sizes in CB, Krumm, and Reichelt in 2000 were inadequate to estimate selection gradients by including all floral traits in a single model. We instead calculated standardized directional selection gradients for all traits except flower number using a series of multiple regression models that each included two independent variables: flower number and one other floral trait. Standardized directional selection differentials were calculated for flower number using univariate regression (Conner, 1988
). We relativized fitness by dividing by mean fitness (Lande and Arnold, 1983
) and standardized each floral trait to a mean = 0 and variance = 1 (Sokal and Rohlf, 1995
). Fitness measures were relativized and floral traits standardized separately for each population and year. We present P values for the selection differentials and gradients both before and after applying the Dunn-Sidak correction (Sokal and Rohlf, 1995
) for multiple tests within each combination of population and year.
The assumption of normality of residual variance was tested using Lilliefors' test (Wilkinson, 1997
). We tested the assumption of homogeneity of residual variance by calculating the Spearman rank correlation between the residuals and relative fitness (Neter et al., 1990
). If the correlation was significant (P < 0.05), we rejected the assumption of homogeneity. We used these methods to test the assumptions of all regression models described in this paper. When necessary, we log-transformed relative fitness to meet the assumptions of regression. A selection differential or gradient calculated using transformed relative fitness is no longer an unbiased point estimate (Lande and Arnold, 1983
). As such, when transformation was necessary, we present the selection differential or gradient from the untransformed model but the P value from the transformed model (Mitchell-Olds and Shaw, 1987
).
We corrected estimates of phenotypic selection for flower number (as in Morgan and Schoen, 1997
; Caruso 2000
, 2001
) because of this trait's overwhelming importance for fitness. In most plant species, including L. cardinalis and L. siphilitica, flower number accounts for a significant amount of the variation in fitness (e.g., Johnston, 1991a
; Conner et al., 1996
; Caruso, 2000
). Correcting estimates of phenotypic selection for flower number is advantageous for two reasons. First, when flower number is included in the model, the variation attributable to this trait is removed from the error term, increasing the power to detect selection. Second, selection for flower number in L. cardinalis and L. siphilitica is consistently stronger than selection for any other trait (Table 2). This selection for flower number could result in strong indirect selection on correlated floral traits. Correcting estimates of phenotypic selection for flower number allows direct selection on floral morphology to be partitioned out from indirect selection via flower number.
|
; Dudley, 1996
; Gilbert et al., 1996
; Caruso, 2000
, 2001
). ANCOVA is the standard method for testing differences in regression coefficients between groups (Sokal and Rohlf, 1995
; Zar, 1999
). By incorporating both population-specific selection estimates and the variation around these estimates, this analysis is also considered to be a powerful method for detecting differences in selection. To detect differences in selection among conspecific populations measured in 2000, separate ANCOVA models were constructed for each species. These models included continuous terms for floral traits, a categorical term coding for population, and floral trait x population terms. The floral trait x population term, if significant, indicates that selection varied spatially.
To detect temporal variation in selection in the CB population, we constructed an ANCOVA model that included a continuous term for the floral trait, a categorical term coding for year, and a floral trait x year term. The floral trait x year term, if significant, indicates that selection varied temporally.
ANCOVA was also used to simultaneously test for spatial and temporal variation in selection on L. siphilitica. The model included a continuous term for the floral trait, categorical terms coding for population and year, and floral trait x categorical factor terms. The floral trait x population term, if significant, indicates that selection varied spatially. The floral trait x year term, if significant, indicates that selection varied temporally. Only L. siphilitica populations that were measured in both 1999 and 2000 were included in the analysis.
Abiotic environment
We used repeated-measures ANOVA to compare soil volumetric water content between L. cardinalis and L. siphilitica populations. In this analysis, species was the between-subjects factor and measurement date was the within-subjects factor.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Spatial variation in selection
Selection on floral morphology did not differ among L. cardinalis populations measured in 2000 (Table 2), although the means for these traits did vary. Corolla lobe length, corolla lobe width, corolla tube length, stigma–nectary distance, and stigma exsertion were all greater in Airport than in CB or Ditch. Plants in Ditch had the widest corollas, followed by those in Airport and CB (Table 3). In contrast to selection on floral morphology, directional selection on flower number did vary among L. cardinalis populations measured in 2000 (Table 2).
|
When we analyzed data from multiple years and populations, we could only detect spatial variation in selection on floral morphology of L. siphilitica. Selection on corolla lobe length (ANCOVA; F1,315 = 9.19, P = 0.003; Fig. 2A), corolla lobe width (F1,315 = 4.58, P = 0.033; Fig. 2B), corolla tube length (F1,315 = 6.42, P = 0.012; Fig. 2C), and stigma–nectary distance (F1,315 = 4.99, P = 0.026; Fig. 2D) differed significantly between CERA and Krumm. Plants with longer and wider corolla lobes, longer corolla tubes, and greater stigma–nectary distance produced more fruits, but only in Krumm. Selection on these traits in CERA was nonsignificant or significantly negative (Fig. 2). With two exceptions (corolla tubes, which were significantly longer in Krumm in 2000 [Table 3], and corolla lobes, which were significantly wider in Krumm in 1999 [Table 4]), comparisons of mean floral traits between populations provide no evidence that differences in selection between CERA and Krumm resulted in character displacement. In contrast to floral morphology, selection on flower number was stronger in CERA than in Krumm (F1,318 = 58.25, P < 0.001).
|
|
|
Abiotic environment
Volumetric soil water content (VWC) was higher in L. cardinalis than L. siphilitica populations (Fig. 4). Averaged across dates, VWC was twice as high in L. cardinalis populations (repeated-measures ANOVA; F1,4 = 27.37, P = 0.006). In addition, there were significant effects of sampling date (F1,4 = 17.37, P = 0.000) and species x sampling date (F1,4 = 7.52, P = 0.001) on VWC.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Although further manipulative experiments would be needed to determine the ecological causes of this variation in selection on floral traits of Lobelia, differences in soil moisture among populations could be an important factor. Lobelia siphilitica may experience greater spatial variation in selection than L. cardinalis (Table 2; Fig. 2) because it occupies a greater variety of moisture environments. All L. cardinalis populations occurred in soils with >50% VWC, where the soil was saturated (Jackson et al., 2000
) and standing water was common (C. M. Caruso, personal observation). In contrast, L. siphilitica populations occupied a greater variety of habitats, from the same saturated soils as L. cardinalis to dry soils that are virtually never saturated (Fig. 4). When data were pooled across populations, coefficients of variation for VWC were greater for L. siphilitica (CV = 42.89%) than L. cardinalis (CV = 31.75%; F435, 391 = 1.43; P < 0.05; Zar, 1999
), as would be expected if L. siphilitica has a greater ecological breadth.
The stronger selection on L. cardinalis in 2000 vs. 1999 may also be associated with variation in soil water availability. The year 1999 was a much wetter year than 2000; 52% more precipitation fell in central Iowa in the 12 mo (November–October) preceding the end of the 1999 flowering season relative to 2000 (data from Todey, 2002
). Temporal variation in selection on floral traits of another species, Platanthera bifolia, has been attributed to variation in rainfall; nectar spurs of P. bifolia were shorter in drier years, but the length of the pollinating hawk moth's proboscis was unchanged, resulting in stronger selection for longer spurs in dry years (Maad, 2000
). The stronger selection on L. cardinalis in 2000 could be the result of a similar mismatch between plant and pollinator. Plants in the CB population produced smaller flowers in 2000 (Tables 3–4), the same year in which there was significantly stronger selection for larger flowers (Fig. 3).
Based on the observed patterns of spatial vs. temporal variation in selection, we predict that L. siphilitica populations should exhibit genetically based differences in floral traits. However, the dimensions of L. siphilitica flowers (Tables 3 and 4) could not be predicted from patterns of selection on floral traits (Table 2; Fig. 2), suggesting that spatial variation in selection on this species has not led to significant phenotypic differentiation. For example, selection on corolla lobe length was stronger in Krumm than in CERA in both 1999 and 2000 (Fig. 2A), but corolla lobes were not significantly longer in Krumm (Tables 3 and 4).
The lack of significant phenotypic differentiation of L. siphilitica in response to spatial variation in selection on floral traits could be the result of the plant's perennial life history, undetected temporal variation in selection, the genetic architecture of floral traits, selection via other fitness components, or gene flow. Our estimates of natural selection for L. siphilitica could be inaccurate because they are based on only a single episode of reproduction in a short-lived perennial species, not lifetime reproductive success (Endler, 1986
). Although we detected little temporal variation in selection on floral traits of L. siphilitica, our data set spanned only 2 yr (1999 and 2000). However, precipitation differed strikingly between these years, indicating that we captured a representative sample of meteorological variation (see earlier data from Todey [2002]
). Negative genetic correlations between flower size and number may constrain the evolution of floral traits in both CERA and Krumm (C. M. Caruso, unpublished data), preventing any adaptive differentiation of floral displays. The combination of positive directional selection on floral traits of L. siphilitica via fruit set and negative selection on these traits via another fitness measure could impede phenotypic differentiation among populations. Although we cannot test this hypothesis for L. siphilitica in the present study, contrasting selection via different fitness components on the same floral trait has been documented in other species (e.g., Solanum carolinense; Elle, 1999
; Elle and Meagher, 2000
). Gene flow between L. siphilitica populations that experience contrasting selective pressures could impede any phenotypic differentiation of floral traits in response to selection (e.g., Endler, 1979
). Because we have not estimated pollen and seed dispersal distances for L. siphilitica, we cannot directly test this hypothesis. However, CERA and Krumm were adjacent to unmeasured L. siphilitica populations (C. M. Caruso, personal observation), suggesting that interpopulation gene flow could occur.
The occurrence of temporal, but not spatial, variation in selection on floral traits of L. cardinalis leads to the prediction that populations of this species should not differ in floral phenotype. Our results provide partial support for this prediction, as only one of six measures of floral morphology differed between CB and Ditch (Table 3). In contrast, plants in the Airport population had consistently larger flowers than those in CB or Ditch (Table 3), which is not consistent with the prediction of a similar phenotype across populations. However, these comparisons may be biased by interpopulation variation in the physical environment. The soils at Airport contain significantly more phosphorus, potassium, and nitrate than at other L. cardinalis populations (C. M. Caruso, unpublished data). In addition, Airport was the only population that occurred under a forest canopy (C. M. Caruso, personal observation). Atmospheric evaporative demand is typically lower under forest canopies than in open areas (Maherali et al., 1997
), suggesting that transpirational water loss is lower in the Airport population. Because both improved plant water status (Carroll et al., 2001
; Elle and Hare, 2002
) and nutrient addition (e.g., Higaki et al., 1992
) can positively influence flower size, the larger flowers in the Airport population may reflect differences in microclimate rather than interpopulation genetic differentiation.
Although L. cardinalis and L. siphilitica are closely related, we documented striking interspecific differences in the pattern of variation in selection on floral traits. Lobelia cardinalis experienced significant temporal variation in selection (Fig. 3), whereas L. siphilitica experienced spatial variation in selection on the same traits (Table 2; Fig. 2). Our data suggest (1) that soil water availability is a potential ecological cause of this variation in selection and (2) that temporal variation in selection has impeded the differentiation of floral phenotypes among L. cardinalis populations. More generally, our results highlight the importance of estimating temporal and spatial variation in natural selection. Although estimates of natural selection have been made for many species, the median number of spatial and temporal replicates in any given study is one (Kingsolver et al., 2001
). Given the potentially contrasting effects of spatial and temporal variation on patterns of phenotypic evolution, such replication is critical for understanding the distribution of phenotypes in nature.
|
FOOTNOTES |
|---|
4 Current address: Department of Botany, University of Guelph, Axelrod Building, 50 Stone Road East, Guelph, ON N1G 2W1 Canada, (phone: 519-824-4120 ext. 52030; FAX: 519-767-1991; carusoc{at}uoguelph.ca
)
5 Current address: Microbiology Doctoral Training Program, University of Wisconsin-Madison, Madison, Wisconsin 53706 USA
6 Current address: Department of Botany and Plant Sciences, University of California-Riverside, Riverside, California 92521 USA
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Beaudoin Yetter R. 1989 The expression of male-sterility in Lobelia siphilitica L. (Campanulaceae): a life history approach. Ph.D. dissertation, Miami University, Oxford, Ohio, USA
Bell G. 1997 Selection, the mechanism of evolution. Chapman and Hall, New York, New York, USA
Bertin R. I. 1982 The ruby-throated hummingbird and its major food plants: ranges, flowering phenology, and migration. Canadian Journal of Zoology 60: 210-219
Campbell D. R. 1991 Effects of floral traits on sequential components of fitness in Ipomopsis aggregata. American Naturalist 137: 713-737[CrossRef][ISI]
Carroll A. B. S. G. Pallardy C. Galen 2001 Drought stress, plant water status, and floral trait expression in fireweed, Epilobium angustifolium (Onagraceae). American Journal of Botany 88: 438-446
Caruso C. M. 2000 Competition for pollination influences selection on floral traits of Ipomopsis aggregata. Evolution 54: 1546-1557[ISI][Medline]
Caruso C. M. 2001 Differential selection on floral traits of Ipomopsis aggregata growing in contrasting environments. Oikos 94: 295-302[CrossRef][ISI]
Conner J. K. 1988 Field measurements of natural and sexual selection in the fungus beetle, Bolitotherus cornutus. Evolution 42: 736-749[CrossRef][ISI]
Conner J. K. 1989 Density-dependent sexual selection in the fungus beetle, Bolitotherus cornutus. Evolution 43: 1378-1386[CrossRef][ISI]
Conner J. K. S. Rush P. Jennetten 1996 Measurements of natural selection on floral traits in wild radish (Raphanus raphanistrum) I. Selection through lifetime female fitness. Evolution 50: 1127-1136[CrossRef][ISI]
Devlin B. 1989 Components of seed and pollen yield of Lobelia cardinalis: variation and correlations. American Journal of Botany 76: 204-214[CrossRef][ISI]
Dudley S. A. 1996 Differing selection on plant physiological traits in response to environmental water availability: a test of adaptive hypotheses. Evolution 50: 92-102
Elle E. 1999 Sex allocation and reproductive success in the andromonoecious perennial Solanum carolinense (Solanaceae). I. Female success. American Journal of Botany 86: 278-286
Elle E. J. D. Hare 2002 Environmentally induced variation in floral traits affects the mating system in Datura wrightii. Functional Ecology 16: 79-88[CrossRef][ISI]
Elle E. T. R. Meagher 2000 Sex allocation and reproductive success in the andromonoecious perennial Solanum carolinense (Solanaceae). II. Paternity and functional gender. American Naturalist 156: 622-636[CrossRef][ISI]
Endler J. A. 1979 Gene flow and life history patterns. Genetics 93: 263-284
Endler J. A. 1986 Natural selection in the wild. Princeton University Press, Princeton, New Jersey, USA
Gilbert F. P. Willmer F. Semida J. Ghazoul S. Zalat 1996 Spatial variation in selection in a plant-pollinator system in the wadis of Sinai, Egypt. Oecologia 108: 479-487[CrossRef][ISI]
Gross J. B. C. Husband S. C. Stewart 1998 Phenotypic selection in a natural population of Impatiens pallida Nutt. (Balsaminaceae). Journal of Evolutionary Biology 11: 589-609[CrossRef][ISI]
Higaki T. J. S. Imamura R. E. Paull 1992 N, P and K rates and leaf tissue standards for optimum Anthurium andraeanum flower production. Hortscience 27: 909-912
Jackson R. B. L. J. Anderson W. T. Pockman 2000 Measuring water availability and uptake in ecosystem studies. In O. E. Sala, R. B. Jackson, H. A. Mooney, and R. W. Howarth [eds.], Methods in ecosystem studies, 199–214. Springer-Verlag, New York, New York, USA
Johnston M. O. 1991a Natural selection on floral traits in two species of Lobelia with different pollinators. Evolution 45: 1468-1479[CrossRef][ISI]
Johnston M. O. 1991b Pollen limitation of female reproduction in Lobelia cardinalis and L. siphilitica. Ecology 72: 1500-1503[CrossRef][ISI]
Johnston M. O. 1992 Effects of cross and self-fertilization on progeny fitness in Lobelia cardinalis and L. siphilitica. Evolution 46: 688-702[CrossRef][ISI]
Kingsolver J. G. H. E. Hoekstra J. M. Hoekstra D. Berrigan S. N. Vignieri C. E. Hill A. Hoang P. Gibert P. Beerli 2001 The strength of phenotypic selection in natural populations. American Naturalist 157: 245-261[CrossRef][ISI]
Lande R. S. J. Arnold 1983 The measurement of selection on correlated characters. Evolution 37: 1210-1226[CrossRef][ISI]
Levins R. 1968 Evolution in changing environments. Princeton University Press, Princeton, New Jersey, USA
Maad J. 2000 Phenotypic selection in hawkmoth-pollinated Platanthera bifolia: targets and fitness surfaces. Evolution 54: 112-123[ISI][Medline]
Maherali H. E. H. DeLucia T. W. Sipe 1997 Hydraulic adjustment of maple saplings to canopy gap formation. Oecologia 112: 472-480[CrossRef][ISI]
Mitchell R. J. 2001 Path analysis: pollination. In S. M. Scheiner and J. Gurevitch [eds.], Design and analysis of ecological experiments, 217–234. Oxford University Press, New York, New York, USA
Mitchell-Olds T. R. G. Shaw 1987 Regression analysis of natural selection: statistical inference and biological interpretation. Evolution 41: 1149-1161[CrossRef][ISI]
Morgan M. T. D. J. Schoen 1997 Selection on reproductive characters: floral morphology in Asclepias syriaca. Heredity 79: 433-441[CrossRef][ISI]
Neter J. W. Wasserman M. H. Kutner 1990 Applied linear statistical models. Irwin, Homewood, Illinois, USA
O'Connell L. M. M. O. Johnston 1998 Male and female pollination success in a deceptive orchid, a selection study. Ecology 79: 1246-1260[CrossRef][ISI]
Pigliucci M. P. Diiorio C. D. Schlichting 1997 Phenotypic plasticity of growth trajectories in two species of Lobelia in response to nutrient availability. Journal of Ecology 85: 265-276[CrossRef]
Pigliucci M. C. D. Schlichting 1995 Ontogenetic reaction norms in Lobelia siphilitica (Lobeliaceae): response to shading. Ecology 76: 2134-2144[CrossRef][ISI]
Reboud X. G. Bell 1997 Experimental evolution in Chlamydomonas III. Evolution of specialist and generalist types in environments that vary in time and space. Heredity 78: 507-514[CrossRef][ISI]
Schemske D. W. C. C. Horvitz 1989 Temporal variation in selection on a floral character. Evolution 43: 461-465[CrossRef][ISI]
Sokal R. R. F. J. Rohlf 1995 Biometry. Freeman, New York, New York, USA
Todey D. 2002 Climatological data selection page. Website, www.agron.iastate.edu/climodat/table.html
Totland O. 2001 Environment-dependent pollen limitation and selection on floral traits in an alpine species. Ecology 82: 2233-2244[CrossRef][ISI]
Underwood A. J. 1997 Experiments in ecology. Cambridge University Press, Cambridge, UK
Wilkinson L. 1997 SYSTAT 7.0: statistics. SPSS, Chicago, Illinois, USA
Zar J. H. 1999 Biostatistical analysis. Prentice Hall, Upper Saddle River, New Jersey, USA
This article has been cited by other articles:
|
|
 |
|
  C. M. Caruso Plasticity of inflorescence traits in Lobelia siphilitica (Lobeliaceae) in response to soil water availability Am. J. Botany, April 1, 2006; 93(4): 531 - 538. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. E. Giblin Variation in floral longevity between populations of Campanula rotundifolia (Campanulaceae) in response to fitness accrual rate manipulation Am. J. Botany, October 1, 2005; 92(10): 1714 - 1722. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Remained significant after Dunn-Sidak correction
