Does inbreeding promote evolutionary reduction of flower size? Experimental evidence from Crepis tectorum (Asteraceae)

? Premise of the study: Small, autogamous flowers have evolved repeatedly in the plant kingdom. While much attention has focused on the mechanisms that promote the shift to autogamy, there is still a paucity of information on the factors that underlie the reduction of flower size so prevalent in selfing lineages. In this study of Crepis tectorum, I examine the role of inbreeding, acting alone or together with selection, in promoting evolutionary reduction of flower size. ? Methods: Experimental crosses were performed to produce progeny populations that differed in inbreeding and (or) selection history. Progenies were grown in two different environments and scored for flower size and other characters. ? Key results: Inbreeding depressed flower and fruit size, but also caused changes in flowering time and the number of heads produced. Despite some inconsistencies in the results for the last progeny generation, the decline in flower size was persistent over generations, consistent across environments, and similar in magnitude to the effects of selection for small flower size and the floral reduction inferred to have taken place during the shift toward autogamy within the study species. The floral size reduction was largely independent of changes in overall vigor, and there was considerable adaptive potential in flower size (measured by sib analyses and parent-offspring comparisons) after inbreeding. ? Conclusions: The results of this study indicate that inbreeding can promote evolutionary reduction of flower size and highlight the close, persistent association between flower and fruit size in the study species.     

Flower-fruit size allometry at three taxonomic levels in Crepis (Asteraceae)

The present investigation utilizes the major axis technique to examine the allometric relationship between flower and fruit size, two developmentally related characters. Regression of log (corolla length) on log (fruit length) using data from 188 species of Crepis , 52 populations of C. tectorum and 40 sibships from a population of C. tectorum demonstrated that flower size shows a decelerating increase with increasing fruit size at all taxonomic levels, with the allometric slope varying from 0.53 to 0.69. The null hypothesis of isometry was rejected in analyses using species or sibships as observations, while the comparison of populations revealed a slope significantly different from 1 only if two outliers were excluded from the analysis. Numerous species and populations have escaped the constraint linking flower and fruit size, including C. tectorum (which has a high ratio of flower to fruit size within the genus) and C. tectorum subsp. pumila (which has a high ratio of flower to fruit size within the species).     

The quantitative genetics of floral trait variation in Lobelia: potential constraints on adaptive evolution

Abstract Although pollinator-mediated natural selection has been measured on many floral traits and in many species, the extent to which selection is constrained from producing optimal floral phenotypes is less frequently studied. In particular, negative correlations between flower size and flower number are hypothesized to be a major constraint on the evolution of floral displays, yet few empirical studies have documented such a trade-off. To determine the potential for genetic constraints on the adaptive evolution of floral displays, I estimated the quantitative genetic basis of floral trait variation in two populations of Lobelia siphilitica . Restricted maximum likelihood (REML) analyses of greenhouse-grown half-sib families were used to estimate genetic variances and covariances for flower number and six measures of flower size. There was significant genetic variation for all seven floral traits in both populations. Flower number was negatively genetically correlated with four measures of flower size in one population and three measures in the other. When the genetic variance-covariance matrices were combined with field estimates of phenotypic selection gradients, the predicted multivariate evolutionary response was less than or opposite in sign to the selection gradient for flower number and five of six measures of flower size, suggesting genetic constraints on the evolution of these traits. More generally, my results indicate that the adaptive evolution of floral displays can be constrained by tradeoffs between flower size and number, as has been assumed by many theoretical models of floral evolution.     

A bryophyte flora of Handa Island Sutherland

ABSTRACT

Background: Hybridization is a key driver of plant speciation, and understanding ongoing hybridization provides insights into the underlying process. We studied a hybrid complex in Mediterranean Cyclamen between common C. repandum and rare C. balearicum on Sardinia.

Aims: We analyzed variation in the frequency and morphology of different floral types to assess the degree of hybridization and analyze the ecological niche to better understand factors that influence the maintenance of hybrid populations.

Methods: We recorded the frequency of the different flower color-types in 45 populations and measured floral traits in 10 of them. We characterized the spatial extent and climatic niche of this hybrid complex. We analyzed floral type variation in four traits (flower size, style and anther length, herkogamy).

Results: Hybrid populations occupy a wide area on the NW coast of the island (~1000km2). Sites dominated by hybrid floral types are ecologically marginal for the common parent. The C. repandum-like type is present in all populations and shows a decrease in style and anther length and herkogamy in populations containing a greater proportion of hybrids.

Conclusions: This hybrid complex provides an ideal situation to examine reproductive isolation due to variation in the ecological niche of different floral types.     

Flower number and floral components in ten angiosperm species: an examination of assumptions about trade-offs in reproductive evolution

A common approach to modelling reproductive evolution in flowering plants includes an implicit assumption that module number and resource allocation per module follow an inverse hyperbolic trade-off. This assumption has not been thoroughly tested. In ten herbaceous and small woody species I examined phenotypic partial correlations between flower number (measured in relation to vegetative biomass) and each of three floral components: pollen number per flower, ovule number per flower, and corolla size. Significantly negative correlations between flower number and at least one of the floral components occurred in four of the ten species. These phenotypic correlations suggest the existence of true (genetically based) trade-offs, because environmental correlations are likely to be positive, but the significant negative relationships are linear except in one case. Thus, evolutionary tradeoffs involving flower number seem likely in some cases, but there is little to indicate that hyperbolic trade-offs are common. The phenotypic patterns investigated here cannot provide definitive answers about the form of trade-offs. Nonetheless, theoretical attention to the potential evolutionary consequences of trade-offs other than the implicit hyperbolic form is needed.     

The genetics of floral divergence and postzygotic barriers between outcrossing and selfing populations of Arenaria uniflora (Caryophyllaceae)

The genetic architecture of floral traits involved in the evolution of self-pollination provides a window into past processes of mating system divergence. In this study, we use two generations of crosses between highly selfing and predominantly outcrossing populations of Arenaria uniflora (Caryophyllaceae) to determine the minimum number, average dominance relationships, and pleiotropic effects of genetic factors involved in floral divergence. Comparison of the F1 and F2 phenotypic means with the expectations of a completely additive model of gene action revealed a primarily additive genetic basis for floral characters associated with mating system variation. The exception was flower life span, which showed partial dominance of the outcrosser phenology. In contrast to similarly divergent species, the substantial differences in flower size between these A. uniflora populations appear to involve relatively few genes of large effect (minimum number of effective factors = 2.2 +/- 2.8 SE). In addition, correlations among traits in the F2 generation indicate that pleiotropy may be an important feature of the genetic architecture of floral evolution in A. uniflora. The evolution of selfing via major modifiers of floral morphology is consistent with other evidence for ecological selection for preemptive self-pollination in A. uniflora. Analyses of the genetic basis of autonomous selfing were complicated by hybrid breakdown in both F1 and F2 generations. Only F1 hybrids showed reductions in female fertility, but about 30% of F1 and F2 hybrids exhibited partial or complete male sterility. Male sterile flowers were characterized by short stamens, reduced petals, and a lack of protandry, as well as indehiscent anthers. This morphological breakdown mimics environmental disruptions of floral development and may result from novel genic interactions in hybrids.     

Gender plasticity in Sagittaria sagittifolia (Alismataceae), a monoecious aquatic species

 Aquatic plants are well known for their high degree of phenotypic plasticity in vegetative structures, particularly leaves. Less well understood is the extent to which their sexuality can be modified by environmental conditions. Here we investigate gender plasticity in the European clonal monoecious aquatic Sagittaria sagittifolia (Alismataceae) to determine how floral sex ratios may vary with plant size and inflorescence order. We sampled two populations from aquatic habitats in East Anglia, U.K. and measured a range of plant attributes including ramet size and the number of female and male flowers per inflorescence. The two populations exhibited similar patterns of phenotypic gender, despite contrasting patterns of total allocation to female and male flower number. Plants produced male-biased floral sex ratios but female flower number increased from the first to the second inflorescence whereas male flower number decreased. Size-dependent gender modification occurred in both populations, but the patterns of allocation to female flower production differed between the two populations. Our results are consistent with the view that monoecy is a sexual strategy that enables plants to adjust female and male allocation in response to changing environmental conditions. Received September 16, 2002; accepted October 23, 2002 Published online: March 20, 2003     

Sex Ratio and Sex‐specific Latitudinal Variation in Floral Characteristics of Gynodioecious Kallstroemia grandiflora (Zygophyllaceae) in Mexico

In species with broad distribution, geographic variation could be associated to particular environmental factors that may affect plant reproduction. In dimorphic species, environmental factors may influence evolutionary processes governing sex ratio variation and floral traits evolution. Kallstroemia grandiflora is an annual insect‐pollinated gynodioecious herb with broad distribution from the Sonoran desert to the semiarid west coast of Mexico. In this study, we explored if female frequency increases in harsh environments and the associated changes in floral morphology in female and hermaphrodite plants. We selected 23 populations along the pacific coast of Mexico where we estimated female frequency and in eight of them we also measured flower size in hermaphrodite and female plants. The average decrease in precipitation was correlated with the increase in female frequency, and a reduction in the flower size of hermaphrodite plants. Our results support the idea that harsh environments may influence female frequency, but also flower size and possibly breeding system evolution.     

北热带喀斯特森林木本植物花性状及其生境分异

植物花性状的多样化是植物长期进化及自然选择的结果, 不同植物种间花性状的变异与生境存在一定的相关性。北热带喀斯特季节性雨林具有生境异质性强、群落结构复杂、特有成分丰富等特点, 分析该森林植物性状的变化特征及其与生境的关联性, 有助于理解物种共存、协同进化过程和对生境的适应, 可为阐明喀斯特植物的生态适应性、理解生物多样性维持机理提供依据。本研究在弄岗15 ha森林动态监测样地木本植物开花相对集中的时间段进行, 记录并分析了21个物种花性状的变化。根据物种空间分布及其与生境的关联特性, 将21个物种分成偏好谷底、山坡、山顶的3种类型, 分析了这3种类型植物的花性状差异; 另外又根据花性状对21个物种进行聚类分析, 探讨了聚类分组结果和根据偏好生境分组结果的异同。结果表明: 花大小、花色鲜艳度均与物种优势度存在显著的负相关关系, 花小而不鲜艳的物种在群落中更有优势, 表现出更好的适应性; 不同偏好生境的植物花色明度有显著差异, 其他的花性状差异不显著; 聚类分组与根据偏好生境的分组有较好的一致性, 反映出物种的花性状变化响应了生境的变化。综上结果, 我们认为北热带喀斯特季节性雨林植物花性状与物种的优势度及生境条件有密切联系, 在该区域中生境对花性状的影响可能比传粉者更为深刻。     

The effects of floral display on pollinator visitation vary among populations ofPhacelia linearis (Hydrophyllaceae)

Summary Individual plants in gynodioecious populations ofPhacelia linearis (Hydrophyllaceae) vary in flower gender, flower size, and flower number. This paper reports the effects of variation in floral display on the visitation behaviour of this species' pollinators (mainly pollen-collecting solitary bees) in several natural and three experimental plant populations, and discusses the results in terms of the consequences for plant fitness. The working hypotheses were: (1) that because female plants do not produce pollen, pollen-collecting insects would visit hermaphrodite plants at a higher rate than female plants and would visit more flowers per hermaphrodite than per female; and (2) that pollinator arrival rate would increase with flower size and flower number, the two main components of visual display. These hypotheses were generally supported, but the effects of floral display on pollinator visitation varied substantially among plant populations. Hermaphrodites received significantly higher rates of pollinator arrivals and significantly higher rates of visits to flowers than did females in all experimental populations. Flower size affected arrival rate and flower visit rate positively in natural populations and in two of the three experimental populations. The flower size effect was significant only among female plants in one experimental population, and only among hermaphrodites in another. The effect of flower number on arrival rate was positive and highly significant in natural populations and in all experimental populations. In two out of three experimental populations, insects visited significantly more flowers per hermaphrodite than per female and visited more flowers on many-flowered plants than on few-flowered plants, but neither effect was detected in the third experimental population. Because seed production is not pollen-limited in this species, variation in pollinator visitation behaviour should mainly affect the male reproductive success of hermaphrodite plants. These findings suggest that pollinator-mediated natural selection for floral display inP. linearis varies in space and time.     

Evolution of floral scent in relation to self-incompatibility and capacity for autonomous self-pollination in the perennial herb Arabis alpina

Background and AimsThe transition from outcrossing to selfing is a frequent evolutionary shift in flowering plants and is predicted to result in reduced allocation to pollinator attraction if plants can self-pollinate autonomously. The evolution of selfing is associated with reduced visual floral signalling in many systems, but effects on floral scent have received less attention. We compared multiple populations of the arctic–alpine herb Arabis alpina (Brassicaceae), and asked whether the transition from self-incompatibility to self-compatibility has been associated with reduced visual and chemical floral signalling. We further examined whether floral signalling differ between self-compatible populations with low and high capacity for autonomous self-pollination, as would be expected if benefits of signalling decrease with reduced dependence on pollinators for pollen transfer.MethodsIn a common garden we documented flower size and floral scent emission rate and composition in eight self-compatible and nine self-incompatible A. alpina populations. These included self-compatible Scandinavian populations with high capacity for autonomous self-pollination, self-compatible populations with low capacity for autonomous self-pollination from France and Spain, and self-incompatible populations from Italy and Greece.Key ResultsThe self-compatible populations produced smaller and less scented flowers than the self-incompatible populations. However, flower size and scent emission rate did not differ between self-compatible populations with high and low capacity for autonomous self-pollination. Floral scent composition differed between self-compatible and self-incompatible populations, but also varied substantially among populations within the two categories.ConclusionsOur study demonstrates extensive variation in floral scent among populations of a geographically widespread species. Contrary to expectation, floral signalling did not differ between self-compatible populations with high and low capacity for autonomous self-pollination, indicating that dependence on pollinator attraction can only partly explain variation in floral signalling. Additional variation may reflect adaptation to other aspects of local environments, genetic drift, or a combination of these processes.     

Comparative genetics of potential prezygotic and postzygotic isolating barriers in a Lycopersicon species cross

I compare the genetic basis of quantitative traits that potentially contribute to pre- and postzygotic isolation between the plant species Solanum lycopersicum (formerly Lycopersicon esculentum) and Solanum habrochaites (formerly Lycopersicon hirsutum), using quantitative trait loci (QTL) mapping in a set of near-isogenic lines. Putative prezygotic isolating traits include flower size, flower shape, stigma exertion, and inflorescence length, that can influence pollinator preferences and/or selfing rates, and therefore gene flow between divergent types. Postzygotic isolating traits are hybrid pollen and seed sterility. Three substantive results emerge from these analyses. First, the genetic basis of floral differentiation appears to be somewhat less complex than the genetic basis of postzygotic hybrid sterility, although these differences are very modest. Second, there is little evidence that traits for floral differentiation are causally or mechanistically associated with hybrid sterility traits in this species cross. Third, there is little evidence that hybrid sterility QTL are more frequently associated with chromosomal centromeric regions, in comparison to floral trait QTL, a prediction of centromeric drive models of hybrid sterility. Although genome-wide associations are not evident in this analysis, several individual chromosomal regions that contain clusters of QTL for both floral and sterility traits, or that indicate hybrid sterility effects at centromere locations, warrant further fine-scale investigation.     

Hormone physiology of pea mutants prevented from flowering by mutations gi or veg1

The veg1 ( vegetative ) mutant in pea ( Pisum sativum L.) does not flower under any circumstances and gi ( gigas ) mutants remain vegetative under certain conditions. gi plants are deficient in production of floral stimulus, whereas veg1 plants lack a response to floral stimulus. During long days in particular, these non-flowering mutant plants eventually enter a stable compact phase characterised by a large reduction in internode length, small leaves and growth of lateral shoots from the upper-stem (aerial) nodes. The first-order laterals in turn produce second-order laterals and so on in a reiterative pattern. The apical bud is reduced in size but continues active growth. Endogenous hormone measurements and gibberellin application studies with gi-1 , gi-2 and veg1 plants indicate that a reduction in gibberellin and perhaps indole-3-acetic acid level may account, at least partially, for the compact aerial shoot phenotype. In the gi-1 mutant, the compact phenotype is rescued by transfer from a 24- to an 8-h photoperiod. We propose that in plants where flowering is prevented by a lack of floral stimulus or an inability to respond, the large reduction in photoperiod gene activity during long days may lead to a reduction in apical sink strength that is manifest in an altered hormone profile and weak apical dominance.     

Native perennial grasses show evolutionary response to Bromus tectorum (cheatgrass) invasion

Invasive species can change selective pressures on native plants by altering biotic and abiotic conditions in invaded habitats. Although invasions can lead to native species extirpation, they may also induce rapid evolutionary changes in remnant native plants. We investigated whether adult plants of five native perennial grasses exhibited trait shifts consistent with evolution in response to invasion by the introduced annual grass Bromus tectorum L. (cheatgrass), and asked how much variation there was among species and populations in the ability to grow successfully with the invader. Three hundred and twenty adult plants were collected from invaded and uninvaded communities from four locations near Reno, Nevada, USA. Each plant was divided in two and transplanted into the greenhouse. One clone was grown with B. tectorum while the other was grown alone, and we measured tolerance (ability to maintain size) and the ability to reduce size of B. tectorum for each plant. Plants from invaded populations consistently had earlier phenology than those from uninvaded populations, and in two out of four sites, invaded populations were more tolerant of B. tectorum competition than uninvaded populations. Poa secunda and one population of E. multisetus had the strongest suppressive effect on B. tectorum, and these two species were the only ones that flowered in competition with B. tectorum. Our study indicates that response to B. tectorum is a function of both location and species identity, with some, but not all, populations of native grasses showing trait shifts consistent with evolution in response to B. tectorum invasion within the Great Basin.     

Floral Display in Narcissus: Variation in Flower Size and Number at the Species, Population, and Individual Levels

Floral display (the size, number, and arrangement of open flowers) influences pollinator visitation to animal-pollinated plants and should be an important determinant of reproductive success. We examined variation in the size and number of open flowers in wild daffodils (Narcissus). Our analysis of published data on 45 taxa showed that flower number varied negatively with flower diameter among Narcissus species, which supports the widespread assumption that there is a trade-off between these traits. In contrast, field measurements indicated a positive relation between flower number and diameter within two populations of Narcissus dubius, and no relation was evident after we controlled for variation in bulb size. The discrepancy between inter- and intraspecific patterns may have occurred because variable resource levels obscure trade-offs when variation in flower size is low (e.g., within species). Size-related increases in floral tube length were half as great as corresponding increases in flower diameter, a result that is consistent with stronger stabilizing selection on tube length. Staggered flowering within N. dubius inflorescences limited the mean number of open flowers to <66% of total flower number, and slow expansion by later opening flowers resulted in significant differences in flower size throughout flowering. Although pollinators preferred large flowers, experimental reductions in flower diameter did not affect seed production. Our results illustrate how the relative importance of the factors influencing floral display can vary among levels of biological organization. Interspecific variation in flower size and number appeared to be constrained by allocation trade-offs, but intraspecific variation in both traits was more greatly influenced by plant resource status. Within plants, the size and number of open flowers reflected the relative age of individual flowers and floral longevity.     

Floral morphometrics and the evolution of sexual dimorphism in Lycium (Solanaceae)

Plants of Lycium californicum, L. exsertum, and L. fremontii produce flowers that are either male-sterile (female) or hermaphroditic, and populations are morphologically gynodioecious. As is commonly found in gynodioecious species, flowers on female plants are smaller than those on hermaphrodites for a number of floral traits. Floral size dimorphism has often been hypothesized to be the result of either a reduction in female flower size that allows reallocation to greater fruit and seed production, or an increase in hermaphroditic flower size due to the increased importance of pollinator attraction and pollen export for hermaphroditic flowers. We provide a test of these two alternatives by measuring 11 floral characters in eight species of Lycium and using a phylogeny to reconstruct the floral size shifts associated with the evolution of gender dimorphism. Our analyses suggest that female flowers are reduced in size relative to the ancestral condition, whereas flowers on hermaphrodites have changed only slightly in size. Female and hermaphroditic flowers have also diverged both from one another and from ancestral cosexual species in several shape characteristics. We expected sexual dimorphism to be similar among the three dimorphic taxa, as gender dimorphism evolved only a single time in the ancestor of the American dimorphic lineage. While the floral sexual dimorphism is broadly similar among the three dimorphic species, there are some species-specific differences. For example, L. exsertum has the greatest floral size dimorphism, whereas L. fremontii had the greatest size-independent dimorphism in pistil characters. To determine the degree to which phylogenetic uncertainty affected reconstruction of ancestral character states, we performed a sensitivity analysis by reconstructing ancestral character states on alternative topologies. We argue that investigations such as this one, that examine floral evolution from an explicitly phylogenetic perspective, provide new insights into the study of the evolution of floral sexual dimorphism.     

Inbreeding and the genetic variance in floral traits of Mimulus guttatus

The additive genetic variance, V(A), is frequently used as a measure of evolutionary potential in natural plant populations. Many plants inbreed to some extent; a notable observation given that random mating is essential to the model that predicts evolutionary change from V(A). With inbreeding, V(A) is not the only relevant component of genetic variation. Several nonadditive components emerge from the combined effects of inbreeding and genetic dominance. An important empirical question is whether these components are quantitatively significant. We use maximum likelihood estimation to extract estimates for V(A) and the nonadditive 'inbreeding components' from an experimental study of the wildflower Mimulus guttatus. The inbreeding components contribute significantly to four of five floral traits, including several measures of flower size and stigma-anther separation. These results indicate that inbreeding will substantially alter the evolutionary response to natural selection on floral characters.     

No evidence for a role of pollinator discrimination in causing selection on flower size through female reproduction

The preference for certain floral phenotypes by flower visiting animals may fuel the evolution of floral traits because variation in flower visitation rates may lead to fitness variation within a population. Here, I examine the importance of flower size for pollinator visitation rate, seed set, and seed mass in two alpine populations of the insect-pollinated herb Ranunculus acris L. during two seasons. There was no pollen limitation of seed set or mass. Pollinators discriminated strongly against flowers experimentally reduced in size. Despite this, there were no signs of any significant impact of flower size on female reproductive success. The results show that although pollinators discriminate strongly among floral phenotypes, this may not always result in female fitness differences within a population because seed set or mass is not limited by pollen availability alone. Probably abiotic environmental constraints prevent plants with high pollinator visitation from capitalizing on the high pollen deposition.     

Herbivory by a Phloem-feeding insect inhibits floral volatile production

There is extensive knowledge on the effects of insect herbivory on volatile emission from vegetative tissue, but little is known about its impact on floral volatiles. We show that herbivory by phloem-feeding aphids inhibits floral volatile emission in white mustard Sinapis alba measured by gas chromatographic analysis of headspace volatiles. The effect of the Brassica specialist aphid Lipaphis erysimi was stronger than the generalist aphid Myzus persicae and feeding by chewing larvae of the moth Plutella xylostella caused no reduction in floral volatile emission. Field observations showed no effect of L. erysimi-mediated floral volatile emission on the total number of flower visits by pollinators. Olfactory bioassays suggested that although two aphid natural enemies could detect aphid inhibition of floral volatiles, their olfactory orientation to infested plants was not disrupted. This is the first demonstration that phloem-feeding herbivory can affect floral volatile emission, and that the outcome of interaction between herbivory and floral chemistry may differ depending on the herbivore's feeding mode and degree of specialisation. The findings provide new insights into interactions between insect herbivores and plant chemistry.     

The role of size and dominance in the feeding behaviour of coexisting hummingbirds

Interspecific competition can strongly influence community structure and limit the distribution and abundance of species. One of the main factors that determine hummingbird community structure is competition for food. The temporal and spatial distribution of nectar has a strong impact on hummingbird assemblages, shaping foraging niches according to hummingbird dominance and foraging strategy. We investigated whether body size and the degree of aggressive dominance influence feeding behaviour of hummingbirds in a temperate forest in northwestern Mexico (El Palmito, Mexico) when winter migrant hummingbirds are present in the community. First, we determined the dominance status of hummingbirds and evaluated the relationship between dominance and body mass, wing disc loading and migratory status. Secondly, we determined how hummingbird species used plant species differently. Thirdly, we examined whether the most dominant hummingbird species defended floral patches with more energy and/or with a larger number of flowers. At each flower patch, hummingbird species, number of hummingbird interactions, feeding time and number of flowers present were recorded. The total number of calories available within each floral patch was also determined. Our results demonstrate that the dominance hierarchy of 13 hummingbird species (migratory and resident) was correlated with body size but not wing disc loading, and that members of the hummingbird community showed a clear separation in resource use (by plant species). Hummingbirds at the top of the dominance hierarchy defended and fed on the best flower patches, defined by the quantity of calories available. Hence, the feeding behaviour of hummingbirds at El Palmito depends on the abundance of plant species used by hummingbirds and on the amount of energy available from each flower patch. Thus, hummingbird body size, aggressive dominance and defence of quality flower patches determines niche partitioning among species.