High reproductive efficiency as an adaptive strategy in competitive environments

Authors

  • Stephen P. Bonser

    Corresponding author
    • Evolution & Ecology Research Centre, School of Biological, Earth and Environmental Sciences, University of New South Wales, Sydney, NSW, Australia
    Search for more papers by this author

Correspondence author. E-mail: s.bonser@unsw.edu.au

Summary

  1. Reproductive efficiency (the efficiency of conversion of resources from vegetative tissue to reproductive output) is a central to our understanding of reproductive allocation and the evolution of reproductive strategies in plants. Plant strategy theory predicts that reproductive efficiency should decrease under competition. Short-lived semelparous species are not predicted to evolve under competition and therefore should not express adaptive responses to the presence of competitors. Long-lived iteroparous species are predicted to delay reproduction in favour of growth and resource acquisition in the presence of competitors. I use life-history theory to advance a prediction that reproductive efficiency increases under competition in both short-lived semelparous and potentially longer-lived iteroparous species.
  2. Contrary to the predictions of plant strategy theory, short-lived semelparous species are frequently observed to live in highly competitive environments. Further, iteroparous species under intense competition may die long before they reach competitive dominance or an optimal size for reproduction.
  3. I surveyed the literature for studies on plant species including measurements of vegetative and reproductive allocation in high and low (or no) competition treatments.
  4. Across species, relative reproductive efficiency (reproductive efficiency under high competition/reproductive efficiency under low competition) significantly increased with increasing competition intensity.
  5. Patterns of allocation to reproduction under competition support the existence of a competitive annual strategy and a reproductive perennial strategy. Under these strategies, short-lived semelparous species and long-lived iteroparous species express high reproductive efficiency under competition as an adaptation to high neighbour density. In addition, some species also expressed patterns of allocation to reproduction consistent with plant strategy theories.
  6. Under this interpretation, I predict that competitive strategies, where plants delay reproduction in competitive environments to gain competitive superiority, are favoured not under intense competition but under modest competition. Including a life-history interpretation in reproductive efficiency under competition provides a much needed predictive framework for strategies of reproduction observed across species.

Introduction

Patterns of allocation of resources to reproduction play a fundamental role in the evolution of life histories and the expression of ecological strategies in plants (see Weiner et al. 2009). The conversion efficiency of vegetative tissues to reproduction (i.e. reproductive efficiency) is central to strategies of reproduction (e.g. Harper & Ogden 1970; Reekie & Bazzaz 1987; Obeso 2002). Reproductive efficiency is a measure of an individual's commitment to reproduction, perhaps at a cost to future growth and reproduction. Here, I develop and evaluate predictions on the expression of reproductive efficiency in competitive environments from the perspectives of plant strategy and life-history theories.

There is a long-standing link between the evolution of life histories and ecological strategies in plants. For example, short-lived, fast growing and highly fecund species are typically found in fertile yet highly disturbed habitats (e.g. Grime 1979). These observations have inspired the development of strategy theories where traits and life histories can be predicted from the environments in which they live. r/K selection (MacArthur & Wilson 1967; Pianka 1970) was an early attempt to predict the evolution of strategies on a gradient of competition. Under r/K selection, individuals in populations below the carrying capacity of the environment experience an environment with relatively low (or no) competition and express a strategy maximizing early life reproduction (r-strategist). In contrast, in populations at or near the carrying capacity experience a high competition environment and favour a strategy of delaying reproduction and increasing allocation to vegetative traits associated with resource acquisition (K-strategist). These early ideas underlying r/K selection theory have been expanded in subsequent theories incorporating axes of stress or impoverishment (e.g. Grime 1979; Taylor, Aarssen & Loehle 1990). While these strategy theories differ in many of their specific predictions, there is general agreement that low competition environments favour high early allocation to reproduction, while high competition environments favour low allocation to reproduction (at least in a given reproductive event) and high allocation to structures associated with resource acquisition.

Allocation to reproduction is a key aspect of plant adaptive strategies across environments. Total lifetime offspring production defines fitness for most organisms. However, many factors can severely limit fecundity but selection can still favour a high commitment to reproduction. For example, both competitively dominant and suppressed plants could express high commitment to reproduction as an adaptive strategy to maximize offspring production in an environment with high plant density and high competition. Here, I use reproductive efficiency (i.e. offspring production per unit vegetative biomass) as a measure of a plant's commitment to reproduction. In its simplest form, reproductive efficiency is a measure of reproductive allocation (e.g. biomass) relative to vegetative allocation (see Harper & Ogden 1970). Reproductive efficiency can also extend to more complex ideas such as the conversion efficiency of resources to reproduction, and conversion efficiency can evolve as an important component of adaptive strategies. The capacity to maximize offspring production for a given vegetative size through efficiently converting resources in vegetative tissues to offspring production may be a fundamental measure of evolutionary success, particularly for suppressed individuals (Bonser & Ladd 2011). In this context, reproductive efficiency is conceptually similar to reproductive economy – that is, traits that ensure offspring production despite limitations to plant size (Aarssen 2008). However, the concept of reproductive efficiency used here extends a plant's commitment to reproduction in benign environments or where there are few limitations on plant size. Reproductive efficiency shares conceptual similarities with other indices of commitment to reproduction such as reproductive effort (often measured as reproductive biomass relative to total biomass) (Harper & Ogden 1970) or fecundity allocation (seed number relative to plant size – see Taylor & Aarssen 1992).

Following the strategy theories outlined above, one can predict that short-lived semelparous species (i.e. annuals and biennials) have high reproductive output and high reproductive efficiency in low competition environments. Total offspring production in their single reproductive event is a key measure of fitness in short-lived semelparous species. High reproductive efficiency will be favoured to maximize the conversion of resources to reproduction and ensure an individual produces as many offspring as it can with the resources available. Several lines of evidence suggest that reproductive efficiency should decrease under competition. For example, resource limitation can sometimes increase the costs of reproduction (e.g. Saulnier & Reekie 1995; but see Jongejans, de Kroon & Berendse 2006; Dostál et al. 2009). Further, competition can suppress the growth of an individual enough that they may struggle to reach a threshold size for reproduction prior to the end of a limited growing season (Aarssen 2008). Reproductive efficiency should decrease under competition where reproduction is size dependent and the plant reaches a size threshold for reproduction but has neither the time nor resources to allocate heavily to reproduction upon reaching this threshold (Bonser & Aarssen 2009). Finally, under plant strategy theory, short-lived semelparous species do not evolve under intense competition (as they are r-selected) and would not likely express effective adaptive responses to the presence of competitors.

For longer-lived iteroparous species, reproductive efficiency in the absence of competition may be relatively high, particularly where plants achieve a size that yields both high fecundity and high reproductive efficiency. Under competition, perennial plants are expected to delay reproduction in favour of allocation to growth and competitive ability (see above). Thus, reproductive efficiency under competition should be low in perennial plants, particularly early in life. Over a lifetime, perennial plants can produce as many offspring as annual plants (Aarssen, Schamp & Pither 2006), but the allocation requirements to be large under competition will likely mean low reproductive efficiency for these plants.

Two observations on how plants respond to competitors suggest that strategy theory does not effectively predict the relationship between reproductive efficiency under competition and the intensity of competition. First, the assumption that short-lived semelparous plants evolve under low or no competition is simplistic. A semelparous annual or biennial life history should evolve where the probability of the adult plant surviving to the next reproductive event is low relative to the probability of offspring surviving the current reproductive event (Charnov & Schaffer 1973; Bell 1976). Disturbances can remove adult plants and open new space for seedlings to establish. While seedlings may start life under low competition, the density of neighbouring plants in these habitats can increase rapidly and competition can feature prominently although the life of many short-lived plants (e.g. Goldberg et al. 2001; Levin & Rees 2002; Schiffers & Tielbörger 2006; Violle, Richarte & Navas 2006; Weinig et al. 2007). I recently demonstrated that annual plants can employ a strategy of increased reproductive efficiency under competition where losses in vegetative size in competition treatments were substantial but losses in fecundity were modest. These results suggest that short-lived semelparous species can evolve under a history of competition and express strategies to minimize the reduction in fitness due to competition (Bonser & Ladd 2011).

Strategy theories such as r/K selection receive a tremendous amount of attention in the literature. This attention can lead to an erroneous conclusion that competition is not an important factor in the evolution of short-lived semelparous plants. Species evolving under solely density-dependent or density-independent selection is probably exceedingly rare. Individuals of any species could experience periods of high and low competition, depending on (among other factors) resource limitation by neighbours, and the potential importance of competition in shaping the strategies of short-lived semelparous plants is probably widely underestimated.

Second, a strategy of delaying reproduction until some optimal size for reproduction is reached, or competitive dominance is realized may not be a viable strategy under intense competition. For example, it is not unusual for iteroparous perennial plants in natural experiments to show little or no growth under competition (e.g. Bonser & Reader 1995; Sammul et al. 2000). There is little evidence to show that competition directly increases mortality. Plants under intense competition can persist as highly suppressed ‘weaklings’ (Aarssen 2008). There is, however, a wealth of evidence demonstrating that mortality is high under competition due to indirect effects such as increased herbivory (e.g. Facelli 1994; Royo & Carson 2008; Chaneton, Mazia & Kitzberger 2010), parasitism or disease (Gilbert 2002; Bell, Freckleton & Lewis 2006). In conditions where competition is sufficiently intense to significantly slow growth, individuals may not live long enough to obtain an optimal size for reproduction or to be competitively dominant, and delaying reproduction could have the severe consequence of reproductive failure. While iteroparous plants have the developmental option of delaying reproduction and persisting in competitive environments, high (or even modest) density-dependent mortality should favour a switch to increased reproduction in competitive environments. High reproductive efficiency in competitively suppressed perennial individuals would allow these plants to produce more offspring over their lives than those plants expressing a strategy of delaying reproduction despite competitive suppression.

These different interpretations of how plants should allocate to reproduction under competition advance contrary predictions on the relationship between reproductive efficiency and the intensity of competition. Under strategy theory, the efficiency of reproduction should be lower under competition, and reproductive efficiency under competition should decrease as the intensity of competition increases. However, under a life-history interpretation, the efficiency of reproduction should be relatively high under competition, and should increase as the intensity of competition increases (Fig. 1). The predictions on reproductive efficiency under strategy theory directly follow predicted patterns of allocation to reproduction. The alternate predictions I make here follow a life-history interpretation but are not explicitly stated under life-history theory. For convenience, I refer to these as predictions made under strategy theory and life-history theory. The opposing nature of the predictions made here is surprising and demonstrates that we currently lack a predictive framework for the relationship between reproductive efficiency and competition. This is a remarkable blind spot in evolutionary ecology, and evaluating and resolving life-history versus plant strategy interpretations are critical in advancing our understanding the nature of plant adaptation to competition.

Figure 1.

Conceptual diagram showing the predicted relationships between relative reproductive efficiency and competition intensity under plant strategy theory and under life-history theory. Relative reproductive efficiency is the reproductive efficiency of plants under high competition relative to those under low or no competition. The dashed line represents the value for relative reproductive efficiency where the efficiency of reproduction under competition is equal to the efficiency of competition in the absence of competition. Competition intensity is a measure of the reduction in plant size in the presence of competitors relative to size in the absence of competition.

Strategies of reproductive efficiency under competition

To test the opposing predictions from plant strategy and life-history theories, I surveyed the literature for studies reporting vegetative and reproductive allocation and/or size across competition treatments. While competition has been a major focus of ecological research, few studies report the effect of competition on both growth and fecundity or reproductive output. Aarssen & Keogh (2002) report only about 3% of competition studies include both of these fundamental variables.

I searched ISI web of science in January 2012 for studies reporting reproductive and vegetative allocation under competition using a combination of keywords including ‘plant competition, neighbour density, reproductive, fecundity, size, reproductive allocation and vegetative allocation’. I also searched for additional studies from references and citations from appropriate studies. To be included in the analysis, studies must have used two competition treatments (e.g. neighbours removed versus neighbours left intact or high versus low neighbour density). Where more than two competition treatments were used, I selected the treatments of highest and lowest competition treatment. There was one exception to these selection criteria (Palmblad 1968) where the highest competition treatment had an extremely high neighbour density and an intermediate competition treatment was used instead. Each study must have recorded some measure of plant size or vegetative size (e.g. total mass, vegetative mass, rosette size, etc.) and reproductive size (e.g. seed number, seed mass, fruit number, inflorescence mass, etc.). Species was used as the unit of replication. Multiple species from the same study were included as independent replicates. In addition, where a given species was included in multiple studies or grown in different sites, then each time the species was included is a replicate. However, where species were grown on an environmental stress gradient in addition to the competition treatments, I only used data from the control or low stress treatments. To ensure the competition treatments were applied throughout the life span of plants, each study had to start with the target plants for each experiments as seedlings. Details of each study used here are included in the Supporting Information (see Appendix S1).

For each of the replicate species, I calculated a relative index of competition intensity:

display math

where Sizelc and Sizehc are measures of plant size under low (or no) competition and under high competition, respectively. Large values of this competition index indicate the effect of competition on the size of a plant is severe. While there are other perhaps more preferable measures of competition intensity, these measures require more detailed aspects of the data, such as the size of the largest plant (e.g. Oksanen, Sammul & Mägi 2006) or the maximum plant performance and the variance of plant performance in the presence and absence of competition (Rees, Childs & Freckleton 2012). These data were not available from the studies I surveyed. The critique of competition indices by Rees, Childs & Freckleton (2012) is primarily on the failure of these indices in testing for differences in competition across stress gradients. My goal here was not to test for a relationship between competition and stress, and the studies included in my data set tended to focus only on competition.

I also calculated an index of relative reproductive efficiency:

display math

where Reff-hc and Reff-lc are the values for reproductive efficiency under high and low (or no) competition, respectively. Relative reproductive efficiency values were log-transformed due to some extremely high values of reproductive efficiency where reproductive efficiency was estimated as seed number (typically a very large value) relative to plant mass (a much smaller value). As these are log values, a positive value of RReff indicates instances where reproductive efficiency is greater under competition than it is under no competition. To avoid autocorrelation problems associated with indices of reproductive allocation (because values of reproductive allocation are sometimes included in both the numerator (reproductive size) and denominator (total size) of the index – see Klinkhamer et al. 1992), all instances where the only values for reproductive size and plant size resulted in autocorrelation were removed from the analysis.

The use of these data has two primary limitations. First, direct comparisons of reproductive efficiency across species are not possible because reproduction has been measured in multiple ways (e.g. total reproductive mass, number of seeds, number of fruits or inflorescence mass). Second, the intensity of competition reported here could be due to the competitive ability of the target plants, or the competitive effect of the competitors. The identity and density of neighbours surrounding the target plants were obviously not controlled across studies, so the impact of competition cannot be interpreted solely as the competitive ability of the target plant. Nevertheless, these data are appropriate for an effective test for a relationship between relative reproductive efficiency and the intensity of competition across species.

Strategies of reproductive efficiency in competitive environments

I found the relative reproductive efficiency (i.e. reproductive efficiency under high competition relative to reproductive efficiency under low or no competition) to increase significantly with the intensity of competition (Fig. 2). Further, for the upper percentiles of the distribution, there is a significantly positive relationship between reproductive efficiency and competition intensity (Quantile regression coefficient for the 0·9 quantile = 0·70, P = 0·016), but there was no significant relationship for the lower percentiles of the distribution (Quantile regression coefficient for the 0·1 quantile = 0·06, P = 0·88). Overall, these results reject strategy theory predictions and support the alternate predictions from life-history theory (see Fig. 1). The distribution of annual, biennial and perennial species on this relationship broadly overlaps and areas of the distribution are not dominated by species of a given life history.

Figure 2.

Scatterplot of relative reproductive effort versus competition intensity across annual (○), biennial (▽) and perennial (□) species. The dashed line represents the value where relative reproductive efficiency is the same under competition as it is in the absence of competition. Pearson product moment correlation statistics are reported on the figure.

The distribution of the expression of reproductive efficiency under competition and the impact of competitors is more complex than a simple continuum of increasing reproductive efficiency with increasing competition intensity. This complex strategy space is represented by three extreme strategies of reproductive efficiency under competition (Fig. 3). The following summary of the observed strategy space is not an attempt to present a new strategy theory. Rather, this summary is intended to highlight the range of strategy space occupied by species as predicted under strategy and life-history theory and to identify the primary axis of variation in strategies observed across species.

Figure 3.

The observed strategy space of relative reproductive efficiency on a gradient of competition intensity across species. The dashed line represents the value for relative reproductive efficiency where the efficiency of reproduction under competition is equal to the efficiency of competition in the absence of competition. The strategy space is highlighted by three extreme strategies: (1) the competitive strategy; (2) the competitive annual and reproductive perennial strategies; and (3) the ruderal strategy (annual plants) and the weak competitive or failed strategy (perennial plants). See text for a full description of these strategies. Most species varied along a continuum of strategy space between strategies 1 and 2.

Low impact of competitors and moderate reproductive efficiency under competition

This strategy space represents a competitive strategy because the presence of competitors has a minimal effect on the growth of target plants. A competitive strategy where reproductive efficiency is similar under high and low competition is predicted under strategy and life-history theories (see Fig. 1). For short-lived semelparous species employing this strategy, vegetative size and growth under competition are likely maintained through a trade-off resulting in reduced allocation to reproduction and in lower reproductive efficiency under competition. For plants of limited life span, a strategy of growth in the presence of competitors is a risky one because slow growth rates limit the time for reproduction and could potentially result in reproductive failure (see Bonser & Aarssen 2009). However, reduction in plant growth here is modest (due to a successful shift to high competitive ability in the target plants and/or low competitive impact of neighbouring plants), and when successful, this gamble could yield high reproductive output under competition.

For longer-lived iteroparous species, the presence of competitors has a minimal impact on the target plants of the experiments, and this space represents a strong competitive strategy. The minimal impact of competition on growth and reproduction could be due to high competitive ability of the target species but could also be due to the low impact of neighbour plants. Where competitors have a minimal impact, these plants should continue to accumulate vegetative size and establish competitive dominance. Competitive dominance of these plants should also permit them to have similar patterns of allocation to reproduction as those plants not under competition.

High impact of competitors and high reproductive efficiency under competition

For short-lived semelparous species, this strategy is consistent with a competitive annual strategy. Under this strategy, high fitness associated with large vegetative size is limited by intense competition. The competitive annual strategy is a response by annual plants to increase reproductive efficiency under competition (see below).

High reproductive efficiency under intense competition is also observed in iteroparous perennial species. This is a surprising result since from the perspective of plant strategy theories perennial species are widely believed to delay reproduction under competition while they accumulate size and competitive dominance, particularly since these are mostly young perennial plants in the 1st or 2nd year of growth. The problem with this strategy is that only a few individuals from a crowded environment will become dominant regardless of the potential life spans of competing individuals. Competition will slow or cease vegetative growth in many or most individuals in a high density community. Thus, in intensely competitive environments, iteroparous plants can express a reproductive perennial strategy, where plants of suboptimal size and poor condition initiate reproduction due to low prospects of surviving to become competitively dominant. Under such a reproductive strategy, plants could achieve relatively high reproductive efficiency by mobilizing a large portion of any stored resources and current photosynthate to reproduction at the cost of reduced future growth and reproduction.

High impact of competitors and low reproductive efficiency under competition

For annual plants, this strategy space represents a ruderal strategy (Grime 1979). Under this strategy, plants do well in the absence of competition but are not equipped to maintain either growth or reproduction under competition. Plants employing a ruderal strategy require intense disturbances that will allow them to rapidly complete a life cycle prior to the onset of competition (including competition from other annual species). Due to the requirement for rapid life cycle completion, biennials and other longer-lived semelparous species are not predicted to (nor are they observed to) occupy this strategy space.

Iteroparous perennial species occupying this strategy space represent a continuum of strategies. At moderate competition intensities, this observed strategy space would be consistent with a weak competitive strategy, where the impact of competitors is not sufficient to induce plants to allocate heavily to reproduction despite their reduction in condition. At high competition intensity, this strategy space represents a failed strategy where plant growth is negligible but plants do not allocate heavily to reproduction either because competition limits the capacity to reproduce or these plants continue with a strategy of achieving competitive dominance.

Notes on the strategy space

Different areas of the strategy space are not occupied by species differing in life history; there is broad overlap in strategy space between annual, biennial and perennial plant species. This does not indicate that short-lived species have roughly the same distribution of competitive ability as long-lived species. Many of these studies were conducted in natural systems where the neighbourhood of competing plants naturally coexist with the target species. As perennial plants probably have higher competitive ability than annual plants, the competitive impact of plants in a perennial community is likely to be greater than the impact of plants in an annual plant community. However, some of these experiments may actually present artificially intense competition. For example, species occupying the failed strategy space (perennials suppressed by competitors but with low reproductive efficiency) may be planted in unnaturally high competitive treatments. The perennial species experiencing the highest combination of competition intensity and the loss of reproductive efficiency under competition (Fig. 2) was a species planted against one or two experimental competitors that may have been vastly superior competitors or planted at unnaturally high densities.

The observed results also do not suggest that reproductive efficiency is similar across life histories. While short-lived semelparous species and long-lived iteroparous species occupy a similar range of relative reproductive efficiencies, a species can have a high value of relative reproductive efficiency so long as reproductive efficiency is high under competition relative to under no competition. In addition, these data cannot be used to make direct comparisons of reproductive efficiency across species because reproductive efficiency was not measured the same way across species. Extensive research on the evolution of life histories demonstrates that allocation to reproduction is greater in semelparous species than it is in iteroparous species (e.g. Pitelka 1977; Silvertown & Dodd 1996; Bonser & Aarssen 2006). The results we present here demonstrate that both short-lived semelparous species and longer-lived iteroparous species often respond to competition by increasing their reproductive efficiency.

Tensions between plant strategy and life-history theories

Competition is recognized as fundamentally important in driving adaptive evolution and controlling populations. Unfortunately, our current framework of plant strategies does not effectively explain the observed reproductive strategies under competition. Plant strategy and life-history theories make generally similar predictions on allocation to reproduction in short-lived semelparous plants and long-lived iteroparous plants, particularly in the absence of competition. However, strategy and life-history theories diverge in their predictions as competition becomes more intense and more pervasive. In particular, strategy theory does not predict the evolution of the competitive annual or the reproductive perennial, both strategies express increased reproductive efficiency in the presence of intense competition.

The evolution of the competitive annual strategy

Following a disturbance, seeds from annual plants either from adult plants occupying the habitat prior to the disturbance or colonizing from other habitats quickly germinate and seedlings establish in an environment relatively free from competition, particularly from adult individuals. In annual plant communities, populations often shift from a low competition environment early in life (e.g. shortly after a disturbance) to a high competition environment later in life as regenerating individuals become large and dense enough to have strong interactions (Schiffers & Tielbörger 2006; Violle, Richarte & Navas 2006). In response to competition, annual plant species (and other short-lived semelparous species) express a competitive annual strategy where plants maximize their reproductive output rather than vegetative growth under competition (Bonser & Ladd 2011).

Competition often reduces the size where reproduction is initiated in annual plants (Kok, McAvoy & Mays 1986; Fone 1989; Bonser & Aarssen 2009; Bonser & Ladd 2011). This response is consistent with a competitive annual strategy where plants should rapidly acquire resources and, if possible, initiate reproduction prior to experiencing intense competition. A strategy of early and intense reproduction will usually result in a loss in size and likely, total fecundity compared to plants grown in the absence of competition. However, rapid increases in neighbour density and limitations for further growth mean that this may be the best strategy to leave some descendants prior to a subsequent disturbance event or the establishment of highly competitive perennial species.

The evolution of the reproductive perennial strategy

‘Things are going to get unimaginably worse, and they are never, ever going to get better again!’ - Kurt Vonnegut. Plants of an iteroparous perennial life history can persist in highly competitive environments but it is not clear whether this strategy is adaptive. Does ecological theory predict that a competitively suppressed individual should eventually become competitively dominant or reach an optimal size for reproduction? Lotka–Volterra models demonstrate that strong competitors tend to drive weak competitors to extinction in two species systems (e.g. Ricklefs & Miller 1999; or any ecology text book). Further, under a competition model of succession, resident species will face increasing competition over time as increasingly competitive species arrive in a habitat (e.g. Ricklefs & Miller 1999). Under many (or even most) circumstances, individuals experiencing diminished growth under competition cannot be expected to eventually overcome their competitors and suppression from competitors will likely get worse, not better. Further, because other factors potentially causing mortality (herbivory, parasitism) tend to be greater under increasing neighbour density, suppressed individuals will often die prior to reaching an optimal size for reproduction.

Size at first reproduction data are limited for perennial species. There is some evidence demonstrating that competition does not affect the size at reproduction (Weiner et al. 2009) or competition can increase the size at reproduction (Nagy & Proctor 1997; Kery, Matthies & Spillmann 2000). Kozlowski (1992) demonstrates that size at reproduction should decrease under conditions of high adult mortality. High reproductive efficiency under intense competition (and limited plant size) observed here (Fig. 2) suggests that size at reproduction often decreases in perennial plants under competition. The observed range of responses of size at reproduction in competitive environments demonstrates that iteroparous perennial plants employ a continuum of strategies from promoting their persistence and eventual dominance in a patch to maximizing their commitment to reproduction even at small sizes prior to their exclusion from a patch. Trees and other very long-lived species have no developmental option for initiating reproduction at very small sizes (or young ages) and many of these species are well known for their persistence strategies (e.g. Loehle 2000). However, many short-lived perennial species (herbaceous perennials) do reproduce at early stages of development and may employ reproduction at small sizes with efficient reproduction to ensure the production of some offspring in a dire competitive situation.

Modest competition from neighbours could potentially hone and improve competitive ability of a strong competitor, one not likely to die prior to reproducing. However, intense competitive interactions should drive a weaker competitor to give up on becoming competitively dominant and allocate heavily to reproduction to avoid the evolutionary dead end of death without sex (see Aarssen 2008). Tracey & Aarssen (2011) advance a model predicting that intense competition should favour reproduction at small sizes, and traits typically associated with high competitive ability (such as large body size) should evolve not under competition but in the absence of competition. These predictions, together with the ideas developed here, suggest that competition as a selection pressure can sometimes, although not always (and perhaps only rarely), select for the evolution of a traditional competitive strategy. This is a surprising, and perhaps counterintuitive, prediction on the conditions promoting the evolution of competitive strategies in plants.

The competitive annual strategy and the reproductive perennial strategy both rely on the capacity for high reproductive efficiency in the face of intense competition. Competition between plants tends to be asymmetric, where large individuals obtain a disproportionately high proportion of available resources (see Weiner 1990). Thus, seedlings are not likely to succeed in a patch already occupied by adult plants, and the removal of adult plants through disturbance (or other factors such as disease) should facilitate seedling establishment. The competitive annual strategy should be a stable strategy for dealing with competitive but frequently disturbed habitats. The success of the reproductive perennial strategy is not as straightforward because disturbances are probably not as strong as that of a feature of perennial assemblages, and patches opened by disturbances would likely be quickly filled by increased growth or spread of remaining adult plants or faster colonizing annual plants. This is a problem for the reproductive perennial throwing the dice on reproducing rather than persisting. Under this strategy, traits that could minimize competition between seedlings and adult plants such as high dispersal ability and or high seed dormancy should be favoured by selection.

Future research is required to establish the competitive conditions (e.g. the relative roles of competitive ability and the competitive effect of neighbouring plants) promoting the expression of high reproductive efficiency under competition. In particular, assessment of the competitive conditions driving the evolution of competitive annuals and reproductive perennials, and the prevalence of these strategies in ecological communities.

The timing of reproduction under competition

A shift toward reproduction at earlier ages under competition could be an important component of high reproductive efficiency in the competitive annual and reproductive perennial strategies. Early reproduction should allow plants to minimize competitive interactions, particularly under conditions where competition is low early in life but increase over time as all individuals in a community increase in size (e.g. Schiffers & Tielbörger 2006). In addition, a strategy of early reproduction under competition could permit more generations (and more total offspring) per unit time than a strategy of delayed reproduction under competition (see Aarssen, Schamp & Pither 2006). Tests for shifts in the age at reproduction under competition are rare, particularly for perennial species. For annuals or biennials, competition or increased neighbour density has been demonstrated both to delay reproduction (Campbell & Snow 2007) and to induce reproduction at earlier ages (Ansari & Daehler 2010). Clearly, we are at an early stage in understanding how competition should affect the timing of reproduction in plants, and this should be a fertile area for future research.

Trade-offs between reproduction and growth

My interpretation of the perhaps unexpected relationship between relative reproductive efficiency and competition intensity rests on trade-offs in allocation between growth and reproduction. For example, increased reproductive efficiency associated with a competitive annual strategy should come at a cost of reduced future growth. The basis of this trade-off could be due to a combination of shorter productive leaf life spans associated with a strategy of higher photosynthesis (e.g. Wright et al. 2004), and a low proportion of resources allocated to later life resource acquisition, particularly in increasingly competitive environments. Similarly, increased reproductive efficiency under competition in perennial plants should come at a cost of decreased future survival as current and stored resources pour into reproductive structures rather than growth and resource allocation structures. The observation that reproductive efficiency increases with increasing competition intensity suggests that plants increasingly forsake any potential future growth and reproduction when it becomes clear that opportunities for future growth and reproduction approach zero (in other words, reproductive efficiency is high where the probably of experiencing the future costs of reproduction is low).

Plants can resorb resources invested in vegetative structures such as leaves. Thus, at the end of the functional life span of leaves, resources can be reallocated to reproduction and reproductive efficiency could be increased through increased resource resorption and reallocation. Intriguingly, resource resorption tends to be greatest in species producing leaves with low specific leaf area (leaves that are tough and/or dense) and long life spans (Kazakou et al. 2007). High photosynthesis versus high resorption strategies for reproductive efficiency imposes a potential constraint on the optimal expression of maximum reproductive efficiency. Where expressing high reproductive efficiency is important, plants of fast life histories likely opt for a strategy of increased photosynthesis and plants of a slow life history likely express a strategy of greater nutrient resorption and reallocation. The role of these trade-offs in the evolution and maintenance of reproductive strategies needs to be tested.

Conclusions

Competition is one of the primary forces defining the struggle for existence and driving adaptive evolution. Plant growth is often limited through resource limitation or other adversities in natural assemblages. Individuals in most vegetation communities experience at least some competition, and competition is often very intense (Tracey & Aarssen 2011). Patterns of allocation to reproduction define fitness in organisms, and reproductive strategies are fundamental in adaptive responses to competitive environments. The current lack of a theoretical framework predicting the observed relationships between reproductive efficiency and competition intensity is a remarkable weakness in evolutionary ecology. I demonstrate that life-history theory can be used to predict high reproductive efficiency under competition in short-lived semelparous species (the competitive annual strategy) and longer-lived iteroparous species (the competitive perennial strategy). Importantly, these strategies do not preclude the expression of other strategies consistent with traditional plant strategy theory such as ruderals and competitors (see Grime 1979), although species tend to express strategies from modest reproductive efficiency under low impact of competition to high reproductive efficiency under high impact of competition. Additional studies estimating reproductive efficiency across competitive environments are required to more effectively test the prediction that competitive strategies evolve under low or modest competition rather than intense competition. Strategy theories have dominated our thinking of plant adaptation and plant strategies under competition. Adding interpretations from life-history theory provides the required framework understanding the nature of adaptation under competition and the evolution of competitive strategies in plants.

Acknowledgements

Joshua Griffiths kindly helped with collecting data from the literature. Angela Moles, Lonnie Aarssen and David Robinson provided helpful comments on earlier versions of this manuscript. This research was supported by an ARC Discovery grant to SPB.

Ancillary