当前位置: X-MOL 学术J. Ecol. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Selection favours high spread and asymmetry of flower opening dates within plant individuals
Journal of Ecology ( IF 5.3 ) Pub Date : 2024-07-22 , DOI: 10.1111/1365-2745.14369
Johan Ehrlén 1, 2 , Alicia Valdés 1, 2
Affiliation  

1 INTRODUCTION

Studies of natural selection on traits expressed repeatedly by individuals usually focus on mean values, although trait distributions within individuals might differ also in other respects. For example, the timing of reproduction of individuals is often described in terms of point estimates, such as onset of breeding in birds, oviposition in insects, leafing out of trees, and onset or peak of flowering. However, in many organisms, the reproduction of individuals consists of multiple events, that is reproductive structures or offspring are produced sequentially, such as the opening of flowers at different times within an individual plant during a reproductive season. As a result, not only the onset of reproduction but also the distribution of reproductive events over the season can vary among individuals, and there might not be a single optimal time for each reproductive event but rather an optimal distribution of reproductive events within individuals. While the importance of reproductive synchrony of individuals within populations has often been acknowledged (Bogdziewicz et al., 2021; Fisogni et al., 2022), the ecological and evolutionary implications of variation in the seasonal distribution of reproductive events within individuals have received far less attention.

The distribution of traits expressed multiple times within individuals can vary regarding the four moments of a distribution, that is the mean, variance, skewness and kurtosis, which describe, the position, degree of spread, asymmetry and weight of the distribution tails, respectively. Figure 1 illustrates how the distribution of flower opening dates over the season (i.e. the flowering schedule) might differ among plant individuals in these four distribution moments. First, individual plants might differ in the mean position of their flowering schedule, some individuals flowering on average earlier than others (Figure 1a). Second, individuals might differ regarding the variance, some individuals having their flowers more spread out in time than others (Figure 1b). Third, skewness of flowering schedules might vary, some individuals producing most flowers during an early peak and remaining flowers during an extended period (Blionis et al., 2001; Forrest & Thomson, 2010; Thomson, 1985), while others produce few flowers during early phases of flowering and have a late peak (Figure 1c). Finally, individual flowering schedules might differ in terms of kurtosis, which refers to the heaviness of the tails of the distribution, a higher kurtosis implying a higher frequency of extreme values (Figure 1d). Some individuals might produce a relatively high number of very early or late-opening flowers, while others produce fewer flowers with extreme opening dates or flowers with less extreme opening dates.

Details are in the caption following the image
FIGURE 1
Open in figure viewerPowerPoint
Theoretical and observed variation in individual flowering schedules, that is the distribution of opening dates of individual flowers. For each of the first four moments of the distribution ([a] mean, [b] variance, [c] skewness and [d] kurtosis) the upper panels show theoretical distributions (i.e. not based on actual data) of opening dates with low, medium and high values of each moment, while keeping the other moments almost constant (difference in other moments <0.1). The lower panels show histograms of the observed values of each moment for the distributions of opening dates for 495 flowering events (i.e. one plant individual flowering in 1 year). The vertical lines on each histogram in the lower panels correspond to the low, medium and high values used in the upper panels.

The different moments of the distribution of reproductive events within individuals can potentially influence fitness in several ways. First, they might influence resource allocation patterns and the ability to provision seeds in different developmental stages that differ in demands. The shape of the flowering schedule is likely to influence within-plant variation in sink strengths and the temporal distribution of total resource demands (Kliber & Eckert, 2004; Medrano et al., 2000). If variance or kurtosis of flower opening dates within individuals are high, and the flowering schedule spread-out or including many extreme opening dates, competition for resources is expected to be lower and the plant might be able to provision more seeds during a reproductive season than would be possible with a more concentrated flowering schedule.

Second, the temporal distribution of reproductive events can affect how individual plants experience the abiotic environment, both directly and via resource availability. For example, if the sensitivity of flowers to adverse weather conditions varies with developmental stage, then the overall effects of variation in weather conditions are likely to be larger in plants with more synchronous flowering than in plants with a more extended flowering. In the same way, the ability to capitalise on beneficial climatic conditions that occur only during a short time window will depend on intra-individual variation in phenology. Increasing variance in opening times of different flowers might thus be regarded as a bet-hedging strategy that reduces fitness variation in response to fluctuating selection, and which is favoured by increasing geometric mean fitness at the expense of average fitness (Childs et al., 2010; Rees & Ellner, 2019; Seger & Brockmann, 1987; Slatkin, 1974). Similarly, more frequent extreme flower opening dates, that is a higher kurtosis, might increase the ability to capitalise on beneficial conditions that occur much earlier or later than the mean flowering date of individuals.

Third, the timing of reproductive events within individuals can influence interactions with other species. Both mutualistic and antagonistic species might respond to the amount and type of within-individual variation, and the net fitness effects and selection mediated by these two groups will depend both on their behavioural responses and on their relative abundance. For example, if a high number of flowers are presented simultaneously (i.e. a strong flowering peak), then this might increase visitation rates to individual flowers if pollinators or pre-dispersal seed predators are disproportionately attracted by large inflorescences (Grindeland et al., 2005; Leimu et al., 2002; Miyake & Sakai, 2005; Molau et al., 1989), or decrease visitation rates if there are satiation effects (Duffield et al., 1993; Elzinga et al., 2007; Honek & Martinkova, 2005; Janzen, 1971; Vanhoenacker et al., 2013). As for interactions with the abiotic environment, a more extended flowering period of individuals, either in terms of a higher variance in flower opening dates or a more frequent production of flowers with extreme opening dates (higher kurtosis), might also serve as a bet-hedging strategy if pollinator availability or antagonist abundance vary unpredictably over the flowering season (Elzinga et al., 2007; Sercu et al., 2021). A right-skewed flowering schedule, with many early opening flowers and a long tail of late-opening flowers, which has been found in several species (Blionis et al., 2001; Forrest & Thomson, 2010; Rathcke & Lacey, 1985; Thomson, 1980, 1985), might increase detectability and attractiveness to pollinators throughout the flowering period (Thomson, 1980). If a right-skewed flowering pattern leads to increased attractiveness to pollinators and higher fruit initiation, this could in turn increase satiation effects on pre-dispersal seed predators with limited egg production. Finally, within-plant variability in timing of reproductive events can affect pollinator movements within plants, and more asynchronous flowering schedules might reduce geitonogamy and promote outcrossing by forcing pollinators to move between plants (Harder & Barrett, 1995).

Given that the temporal distribution of reproductive events might vary among individuals in several respects, and that the optimal distribution of reproductive events depends on both resource allocation patterns and interactions with the abiotic and biotic environment, we expect that natural selection potentially can target all aspects of this variation. In spite of this, most studies of phenotypic selection have used point estimates, such as mean phenology or dates of first reproductive events, while only a few have examined selection on other aspects of the distribution of reproductive events within individuals (Forrest & Thomson, 2010; March-Salas et al., 2021; Sercu et al., 2021), and none has assessed selection on multiple moments of this distribution. The most commonly used measure of flowering phenology, first flowering date, can be influenced by all moments of the distribution of flower opening dates, but this has largely been overlooked on studies of selection on flowering phenology (Inouye et al., 2019). Herrera (2009) proposed an extension of phenotypic selection models that included not only trait means but also other aspects of within-plant trait variation. In this study, we applied this extension to assess phenotypic selection on flowering schedules in the perennial herb Lathyrus vernus, in terms of the mean, variance, skewness and kurtosis of within-individual distributions of flower opening dates. Using data from recordings of 5287 individual flowers belonging to 495 flowering events over 3 years, we quantified mean, variance, skewness and kurtosis of flowering schedules (i.e. the first four moments of the distribution of opening dates of individual flowers) and addressed the following questions: (1) How much of the variation in flowering phenology can be attributed to within-individual variation among flowers on the same plants versus among-individual variation among plants within the population? (2) Is there phenotypic selection on within-individual variation in flowering schedules, and if so, does selection vary among years? (3) How is selection related to effects of within-individual variation in flowering schedules on two fitness components; fruit set and the proportion of seeds escaping predation?



中文翻译:


选择有利于植物个体内开花日期的高传播和不对称性


 1 引言


对个体重复表达的性状的自然选择的研究通常侧重于平均值,尽管个体内部的性状分布在其他方面也可能有所不同。例如,个体的繁殖时间通常用点估计来描述,例如鸟类繁殖的开始、昆虫的产卵、树木的叶子以及开花的开始或高峰。然而,在许多生物体中,个体的繁殖由多个事件组成,即生殖结构或后代是按顺序产生的,例如在生殖季节单个植物内的不同时间开花。因此,不仅繁殖的开始,而且繁殖事件在季节中的分布也会因个体而异,并且每个繁殖事件可能没有一个最佳时间,而是个体内部生殖事件的最佳分布。虽然种群内个体生殖同步的重要性通常已得到承认(Bogdziewicz et al., 2021;Fisogni et al., 2022),个体体内生殖事件季节性分布变化的生态和进化影响受到的关注要少得多。


个体内多次表达的性状分布可能因分布的四个时刻而异,即均值、方差、偏度和峰度,它们分别描述了分布尾部的位置、传播程度、不对称性和权重。图 1 说明了在这四个分布时刻,花朵在季节中的开放日期分布(即开花时间表)在植物个体之间可能如何不同。首先,单个植物的开花时间表的平均位置可能不同,一些个体平均比其他个体开花早(图 1a)。其次,个体在方差方面可能有所不同,一些个体的花朵在时间上比其他个体更分散(图 1b)。第三,开花时间表的偏度可能会有所不同,一些个体在早期高峰期开出最多的花,而在较长的时间内开出剩余的花(Blionis等人,2001;Forrest & Thomson, 2010;Thomson, 1985),而其他花在开花的早期阶段开花很少,高峰期较晚(图 1c)。最后,各个开花时间表在峰度方面可能有所不同,峰度是指分布尾部的重量,较高的峰度意味着极值的频率较高(图 1d)。一些个体可能会产生相对数量相对较多的非常早或非常晚开放的花朵,而另一些个体可能会产生较少的开放日期或不太极端开放的花朵。

Details are in the caption following the image
 图 1

在图窗查看器PowerPoint 中打开

单个开花时间表的理论和观测变化,即单个花卉的开放日期分布。对于分布的前四个时刻([a] 均值、[b] 方差、[c] 偏度和 [d] 峰度),上面板显示了开盘日期的理论分布(即不基于实际数据),每个时刻都有低、中和高值,同时保持其他时刻几乎不变(其他时刻的差值 <0.1)。下面的面板显示了 495 次开花事件(即 1 年内一株植物单独开花)的开放日期分布的每个时刻的观测值的直方图。下部面板中每个直方图上的垂直线对应于上部面板中使用的低、中和高值。


个体体内生殖事件分布的不同时刻可能会以多种方式影响健康状况。首先,它们可能会影响资源分配模式和在不同需求不同的开发阶段提供种子的能力。开花时间表的形状可能会影响植物内部汇强度的变化和总资源需求的时间分布(Kliber & Eckert, 2004;Medrano et al., 2000)。如果个体内部开花日期的差异或峰度很高,并且开花时间表分散或包括许多极端的开花时间,则预计对资源的竞争会较低,并且植物可能能够在繁殖季节提供比更集中的开花时间表更多的种子。


其次,生殖事件的时间分布会影响单个植物如何体验非生物环境,无论是直接的还是通过资源可用性。例如,如果花朵对恶劣天气条件的敏感性随发育阶段而变化,那么天气条件变化的总体影响在同步开花较多的植物中可能比在开花时间较长的植物中更大。同样,利用仅在短时间内出现的有利气候条件的能力将取决于物候的个体内部变化。因此,增加不同花朵开放时间的差异可以被视为一种赌注对冲策略,它减少了对波动选择的响应的适应性变化,并且通过以牺牲平均适应性为代价增加几何平均适应性来支持这种策略(Childs 等人,2010 年;Rees & Ellner, 2019;Seger & Brockmann, 1987;Slatkin, 1974)。同样,更频繁的极端开花日期,即更高的峰度,可能会增加利用比个体平均开花日期早或晚发生的有利条件的能力。

Third, the timing of reproductive events within individuals can influence interactions with other species. Both mutualistic and antagonistic species might respond to the amount and type of within-individual variation, and the net fitness effects and selection mediated by these two groups will depend both on their behavioural responses and on their relative abundance. For example, if a high number of flowers are presented simultaneously (i.e. a strong flowering peak), then this might increase visitation rates to individual flowers if pollinators or pre-dispersal seed predators are disproportionately attracted by large inflorescences (Grindeland et al., 2005; Leimu et al., 2002; Miyake & Sakai, 2005; Molau et al., 1989), or decrease visitation rates if there are satiation effects (Duffield et al., 1993; Elzinga et al., 2007; Honek & Martinkova, 2005; Janzen, 1971; Vanhoenacker et al., 2013). As for interactions with the abiotic environment, a more extended flowering period of individuals, either in terms of a higher variance in flower opening dates or a more frequent production of flowers with extreme opening dates (higher kurtosis), might also serve as a bet-hedging strategy if pollinator availability or antagonist abundance vary unpredictably over the flowering season (Elzinga et al., 2007; Sercu et al., 2021). A right-skewed flowering schedule, with many early opening flowers and a long tail of late-opening flowers, which has been found in several species (Blionis et al., 2001; Forrest & Thomson, 2010; Rathcke & Lacey, 1985; Thomson, 1980, 1985), might increase detectability and attractiveness to pollinators throughout the flowering period (Thomson, 1980). If a right-skewed flowering pattern leads to increased attractiveness to pollinators and higher fruit initiation, this could in turn increase satiation effects on pre-dispersal seed predators with limited egg production. Finally, within-plant variability in timing of reproductive events can affect pollinator movements within plants, and more asynchronous flowering schedules might reduce geitonogamy and promote outcrossing by forcing pollinators to move between plants (Harder & Barrett, 1995). 

Given that the temporal distribution of reproductive events might vary among individuals in several respects, and that the optimal distribution of reproductive events depends on both resource allocation patterns and interactions with the abiotic and biotic environment, we expect that natural selection potentially can target all aspects of this variation. In spite of this, most studies of phenotypic selection have used point estimates, such as mean phenology or dates of first reproductive events, while only a few have examined selection on other aspects of the distribution of reproductive events within individuals (Forrest & Thomson, 2010; March-Salas et al., 2021; Sercu et al., 2021), and none has assessed selection on multiple moments of this distribution. The most commonly used measure of flowering phenology, first flowering date, can be influenced by all moments of the distribution of flower opening dates, but this has largely been overlooked on studies of selection on flowering phenology (Inouye et al., 2019). Herrera (2009) proposed an extension of phenotypic selection models that included not only trait means but also other aspects of within-plant trait variation. In this study, we applied this extension to assess phenotypic selection on flowering schedules in the perennial herb Lathyrus vernus, in terms of the mean, variance, skewness and kurtosis of within-individual distributions of flower opening dates. Using data from recordings of 5287 individual flowers belonging to 495 flowering events over 3 years, we quantified mean, variance, skewness and kurtosis of flowering schedules (i.e. the first four moments of the distribution of opening dates of individual flowers) and addressed the following questions: (1) How much of the variation in flowering phenology can be attributed to within-individual variation among flowers on the same plants versus among-individual variation among plants within the population? (2) Is there phenotypic selection on within-individual variation in flowering schedules, and if so, does selection vary among years? (3) How is selection related to effects of within-individual variation in flowering schedules on two fitness components; fruit set and the proportion of seeds escaping predation? 

更新日期:2024-07-22
down
wechat
bug