CONCEPTS & SYNTHESIS EMPHASIZING NEW IDEAS TO STIMULATE RESEARCH IN ECOLOGY
Ecology, 88(2), 2007, pp. 271–281 Ó 2007 by the Ecological Society of America
EXPANDING THE LIMITS OF THE POLLEN-LIMITATION CONCEPT: EFFECTS OF POLLEN QUANTITY AND QUALITY MARCELO A. AIZEN1,3
LAWRENCE D. HARDER2
AND
1
Laboratorio Ecotono, Universidad Nacional del Comahue, Centro Regional Bariloche, Quintral 1250, 8400 San Carlos de Bariloche, Rio Negro, Argentina 2 Department of Biological Sciences, University of Calgary, Calgary, Alberta T2N 1N4 Canada
Abstract. Pollination commonly limits seed production, as addition of pollen to stigmas often increases fecundity. This response is usually interpreted as evidence that plants’ stigmas receive too few pollen grains to maximize ovule fertilization (quantity limitation); however, many genetic studies demonstrate that poor-quality pollen can also reduce seed production (quality limitation). We explore both aspects of pollen limitation theoretically with a dose– response model that incorporates a saturating negative-exponential relation of seed production to pollen receipt. This relation depends on aspects of ovule production, pollen import, pollen–pistil interactions and seed development, all of which can contribute to pollen limitation. Our model reveals that quantity limitation is restricted to the lowest range of pollen receipt, for which siring success per pollen grain is high, whereas quality limitation acts throughout the range of pollen receipt if plants do not import the highest-quality pollen. In addition to pollinator availability and efficiency, quantity limitation is governed by all postpollination aspects of seed production. In contrast, quality limitation depends on the difference in survival of embryos sired by naturally delivered pollen vs. by pollen of maximal quality. We briefly illustrate the distinction between these two components of pollen limitation with results from the mistletoe Tristerix corymbosus. Our model also shows that the standard pollen-supplementation technique neither estimates the total intensity of pollen limitation nor distinguishes between its quantity and quality components. As an alternative, we propose a methodological protocol that requires both measurement of seed production following excess pollination with only outcross pollen and quantification of the dose–response relation of seed output to pollen receipt. This method estimates both the total extent of pollen limitation and its two components. Finally, we consider the influences on quantity and quality limitation, which reveals that quantity limitation probably occurs much less often than has been inferred from pollen-supplementation experiments. These interpretations suggest that an expanded perspective that recognizes the fecundity consequences of pollination with poor-quality pollen would promote ecological understanding of pollen limitation. Key words: plant reproduction; pollen limitation; pollen quality; pollination; seed production; supplemental pollination; Tristerix corymbosus.
INTRODUCTION Outcrossing plants mate only with the assistance of pollen vectors, so that the abundance and efficiency of vectors determine mating success. Recent surveys for many species suggest that insufficient pollen receipt may commonly compromise seed production in plant populations (Burd 1994, Larson and Barrett 2000, Ashman et Manuscript received 14 June 2006; revised 21 August 2006; accepted 29 August 2006. Corresponding Editor: M. D. Eubanks. 3 E-mail:
[email protected]
al. 2004, Knight et al. 2005). For example, the most recent survey considered 482 studies of fruit production and concluded that 63% of species exhibit pollen limitation at some sites, or during some years (Knight et al. 2005). If this finding is representative, pollen limitation may commonly hamper the ability of plants to realize their reproductive capacity. Such limitation could affect individual and population performance, although the frequency with which seed production, rather than seedling establishment, limits plant recruitment is poorly understood. Furthermore, theory indicates that pollen limitation can determine a minimum
271
CONCEPTS & SYNTHESIS
272
MARCELO A. AIZEN AND LAWRENCE D. HARDER
plant density for population persistence (Morgan et al. 2005). Such effects on reproductive output bear significant and diverse consequences for natural populations, including those subject to human interference, and for the productivity of outcrossing, agricultural crops (Silvertown et al. 1993, Parker 1997, Fre´ville et al. 2004, Chacoff and Aizen 2006). Pollen limitation comprises two components: quantity limitation and quality limitation. Traditionally, pollen limitation has been characterized as the consequence of plants receiving too few pollen grains to fertilize all of their ovules (quantity limitation). For example, Knight et al. (2005: 468) stated that ‘‘(p)ollen limitation occurs when plants produce fewer fruits and/or seeds than they would with adequate pollen receipt.’’ However, the large literature on inbreeding depression demonstrates that pollen quality effects associated with both self-fertilization and mating between related plants (biparental inbreeding: Griffin and Eckert 2003, Herlihy and Eckert 2004) can also reduce seed production (Charlesworth and Charlesworth 1987), probably because embryos homozygous for deleterious alleles die during development (see Plate 1). In addition, self-pollen can interfere with the performance of cross-pollen in self-incompatible species (Ramsey and Vaughton 2000, Kawagoe and Suzuki 2005) and self-pollen tubes can disable ovules in species with ovarian, or late-acting, self-incompatibility, even though fertilization does not occur (Sage et al. 1994). Thus, Husband and Schemske (1996) found that, compared to cross-pollination, self-pollination reduced seed production in 62 self-compatible plant species or populations by an average of 20%, with a maximum reduction of 87% (also see Lloyd and Schoen 1992). Such ‘‘quality limitation’’ may occur commonly, because flowers often receive self-pollen, in addition to crosspollen, as a result of either autonomous processes or the action of pollen vectors. Futhermore, cross-pollination from very distant plants may reduce seed production through outbreeding depression (Waser and Price 1991). In addition to factors related to the intrinsic performance of embryos, plants may not mature all embryos into seeds because of maternal choice among seeds sired by different mates (Obeso 2004). Finally, plants may commonly receive heterospecific pollen due to indiscriminate pollen vectors, which can interfere with the post-pollination performance of conspecific pollen (Murphy 2000, Brown et al. 2002), or result in unviable hybrid fertilizations (Grant 1981). Despite recognition of these diverse quality effects on seed production (Ramsey 1995, Herrera 2000, Ramsey and Vaughton 2000, Anderson and Hill 2002, Finer and Morgan 2003, Ashman et al. 2004), their implications for pollen limitation have not been considered thoroughly. Quality effects could confound estimates of the importance of insufficient pollen import as a limit on seed production, because the common technique for quantifying pollen limitation may often involve higher quality pollen than plants receive naturally, on average.
Ecology, Vol. 88, No. 2
This protocol entails the addition of abundant (‘‘unlimited’’) outcross pollen to the stigmas of some or all of a plant’s flowers, which have previously been exposed to natural (open) pollination (reviewed by Ashman et al. [2004]). Fruit and/or seed production by these ‘‘supplemented’’ flowers is compared with that of non-supplemented, open-pollinated flowers on the same or other plants, and any fertility deficit by open-pollinated flowers is interpreted as quantitative pollen limitation. However, by adding pollen of possibly higher genetic quality, pollen supplementation may also overcome some qualitative pollen limitation, thereby inflating the estimate of quantitative limitation. We propose that qualitative factors limit seed production more commonly than is usually acknowledged. Because many plant populations are genetically structured (Heywood 1991) and pollen vectors typically disperse pollen locally (Harder and Barrett 1996), the deposition of self-pollen, or more generally of pollen from genetically related donor plants, is probably the rule rather than the exception, despite many floral mechanisms that limit selfing and/or promote outcrossing. In addition, inbreeding effects of pollen quality can extend through the life cycle of progeny and beyond (Husband and Schemske 1996), emphasizing the importance of recognizing and quantifying the impact of quality limitation. In this paper, we present a simple dose–response model that provides a conceptual framework for distinguishing the effects of quantitative and qualitative limitation and for estimating their relative contributions to overall pollen limitation. This model also illustrates that pollen supplementation can neither distinguish between these quantitative and qualitative components, nor estimate the total magnitude of pollen limitation. This conclusion raises important questions about whether literature surveys of supplementation studies accurately represent the frequency with which receipt of insufficient pollen limits seed production. THE RELATIVE MAGNITUDES OF QUANTITY AND QUALITY LIMITATION From the maternal perspective, a flower’s seed production involves two post-pollination phases, fertilization and seed development, both of which may depend on the amount and quality of pollen received by the stigma. In this context, ‘‘pollen quality’’ refers to both the ability of a pollen grain on a stigma to fertilize an ovule and the survival probability of the resulting embryo. Suppose that a pistil contains O ovules and that its stigma receives P pollen grains, which may differ in quality, resulting in p ¼ P/O pollen grains per ovule. Germination failure and attrition of pollen tubes in the style reduce the number of male gametophytes in the pistil, so that pollen tubes from only a proportion, b, of the original P pollen grains reach the ovary and compete for fertilization. If the pollen tubes in the ovary distribute randomly among ovules, the number of pollen tubes that can fertilize each ovule will follow a Poisson
February 2007
POLLEN LIMITATION
distribution with a mean of bp. Consequently, if one pollen tube is sufficient to fertilize an ovule, the probability that an ovule is not fertilized (i.e., no pollen tube) is ebp, so the probability that it is fertilized is 1 ebp. Thus a total of (1 ebp )O ovules will be fertilized. If a fraction d of fertilized ovules mature into seeds, for a given mixture of pollen, the flower’s total expected seed production is ð1Þ S ¼ dO 1 ebp : Other studies have modeled this dose–response relation using total pollen receipt (P) rather than pollen receipt per ovule ( p), 0 S ¼ dO 1 eb P :
The relative effects of pollen quantity and quality on seed production can be illustrated by considering a pistil that receives too few pollen grains of low quality (PL) to fertilize all ovules. In this case, there are pL pollen grains per ovule (e.g., solid symbol in Fig. 1b) and the seed production increases with pollen receipt as determined by parameters bL and dL (Fig. 1b, dotted curve), rather than according to the dose–response relation that would result if the stigma received the highest-quality pollen, determined by bH and dH (Fig. 1b, solid curve). Because of quantity limitation, OebL pL ovules remain unfertilized, so that kN ¼ dL OebL pL
ð2Þ
fewer seeds are produced than if pollen vectors had delivered more pollen (quantity limitation: dashed arrow in Fig. 1b). In addition, the pistil could have produced kQ ¼ ðdH dL ÞO
ð3Þ
more seeds if its stigma had been saturated only with high-quality pollen (quality limitation: solid arrow in Fig. 1b). The proportion of total pollen limitation associated with pollen quality (kQ/[kQ þ kN]) ranges from a minimum of 1 (dL/dH), when stigmas receive no pollen, to 1, when stigmas receive enough low-quality pollen to fertilize all ovules. Therefore, the effect of pollen quality on the proximate limitation of seed production depends entirely on the differential success of embryos fertilized by low- and high-quality pollen during seed development. Note that quantity limitation can be alleviated by increased pollen import, whereas quality limitation cannot. EMPIRICAL EXAMPLE
OF
QUALITY LIMITATION
We illustrate the effects of pollen quality on pollen limitation with observations of Tristerix corymbosus (L.) Kuijt (Loranthaceae), a hummingbird-pollinated mistletoe from the southern Andes. Like other loranthaceous mistletoes, this species produces single-seeded fruits, so that fruit set and seed set are equivalent. Reanalysis of data collected during 1996 (see Aizen 2005) demonstrates that hand-pollination with pure self-pollen (Fig. 2, gray symbol) causes 20% lower fruit set than that with pure cross pollen (Fig. 2, black symbol: generalized linear model, G1 ¼ 3.81, P ¼ 0.05). Open-pollinated flowers on the same plants (open symbol) also set fewer fruits than exclusively cross-pollinated flowers (G1 ¼ 6.03, P , 0.025), but not than self-pollinated flowers (G1 ¼ 0.22, P . 0.5). Thus, open-pollinated flowers do not realize the reproductive potential demonstrated by cross-pollinated flowers. At the same site during 1997 and 1998, Aizen (2003) quantified pollen tubes in the stigmas and fruit set for individual open-pollinated flowers on 20 plants (Fig. 2, gray dots), which allowed us to characterize the dose– response relation described by Eq. 1 using nonlinear regression (see Representative assessment of pollen
CONCEPTS & SYNTHESIS
where b 0 ¼ b/O and O is average ovule production (e.g., Waser and Price 1991, Mitchell 1997, Aizen and Basilio 1998). We recommend the approach depicted in Eq. 1, when possible, because by accounting for variation in ovule number among flowers it should reduce variation in the observed dose–response relation, increasing the reliability of statistical estimates of b and d. Eq. 1 suggests that maximal seed production (dO) can be limited by ovule production (O), or by abortion due to genetic death, maternal choice, or sibling competition for limited maternal resources (through d ). The theory that we develop incorporates no assumptions about the association of b and d, but we expect that they often covary positively (e.g., Waser and Price 1991). For example, deleterious alleles that reduce the performance of self-pollen during germination, pollen-tube growth, and fertilization should also reduce the performance of selfed embryos during seed development (Charlesworth and Charlesworth 1987). In contrast, our model assumes that the seed set by individual flowers is not affected by the fecundity of other flowers on the same plant, even though reallocation of maternal resources among flowers is known to influence fruit production, including the outcome of pollen-supplementation experiments (reviewed by Ashman et al. [2004]). Fig. 1a illustrates the relation of seed production to pollen receipt for four pollen mixtures that differ in the proportion of high-quality pollen (i.e., different values of b and d ), with the dotted line representing the mixture with lowest average quality (perhaps including much self-pollen) and the solid line representing the highestquality mixture (perhaps including only cross-pollen). In all cases, the ovary contains 50 ovules, so that the differences in maximal seed production, dO, result from differences in the proportion of fertilized ovules that develop into seeds. In addition, the asymptotic seed production is approached more quickly with increasing pollen receipt for high-quality pollen than for lowquality pollen, because of higher values of both b and d (dS/dp ¼ bdOebp ). As a result, more low- than highquality pollen grains are required to fertilize the same proportion of ovules, and fewer seeds result.
273
MARCELO A. AIZEN AND LAWRENCE D. HARDER
Ecology, Vol. 88, No. 2
CONCEPTS & SYNTHESIS
274
FIG. 1. Relation of seed production to pollen receipt per ovule (P, number of pollen grains; O, number of ovules) for pollen mixtures that differ in the proportion of high-quality pollen for ovaries, based on Eq. 1. (a) Seed production by ovaries that contain 50 ovules and receive pollen mixtures differ in the relation of fertilization to pollen receipt (b [the proportion of the original pollen grains that reach the ovary] ¼ 0.10, 0.15, 0.20, and 0.25 for the dotted, dashed, dashed-dotted, and solid lines, respectively) and the fraction of fertilized ovules that develop into seeds (d ¼ 0.4, 0.5, 0.6, and 0.7 for the dotted, dashed, dashed-dotted, and solid lines, respectively). (b) Conceptual representation of components of pollen limitation of seed production for an ovary that receives lowquality pollen (closed symbol), including quantity limitation (kN) and quality limitation (kQ). In (b) the solid and dotted curves depict the dose–response curves for increased receipt of the highest-quality pollen (determined by bH and dH) and low-quality pollen (determined by bL and dL), respectively, whereas the solid and dotted horizontal lines represent the corresponding maximal possible seed production for each pollen type. The open symbol in (b) represents the seed production that would have resulted had the flower received the same number of high-, rather than low-quality pollen grains. The solid vertical arrow shows quality limitation (kL); the dashed arrow shows quantity limitation (kN).
limitation, below, for methods). Because the available observations involve pollen tubes in stigmas, rather than pollen grains, this analysis overestimates b, the proportion of the original pollen grains that reach the ovary; asymptotic fruit set (d ) is estimated accurately, so that consideration of pollen tubes does not affect our interpretations (see Fig. 1b). The analysis that we present here (Fig. 2, solid curve) considers the natural pollination data for both years, because a regression model with common parameter estimates for both years
was associated with a lower Akaike’s information criterion (AIC ¼ 1481.1) than a model with separate estimates (AIC ¼ 1484.7). The similar dose–response relations for the two years (and at a second site; results not shown) indicate consistent pollination conditions, which we assume also prevailed during 1996. The asymptotic fruit set estimated for the dose– response relation strikingly resembles fruit set by both open- and self-pollinated flowers during 1996 (Fig. 2). This correspondence suggests that open-pollinated flow-
February 2007
POLLEN LIMITATION
275
CONCEPTS & SYNTHESIS
FIG. 2. Dose–response relation of fruit set by Tristerix corymbosus to varying pollination intensity. (a) The histogram represents the frequency distribution of the number of flower samples with different numbers of pollen tubes and illustrates the number of observations (i.e., flowers) for each pollen-receipt category (based on Aizen [2003]). (b) Relation of the proportion of flowers setting fruit (fruit set) to the number of pollen tubes in the stigmas exposed to different pollination conditions at Penı´ nsula de San Pedro, Nahuel Huapi National Park, Argentina. The data (means 6 SE) represent the outcomes of open pollination (open circle), pure self-pollination (gray square), and pure cross-pollination (black diamond) during 1996, with observations of fruit set and pollen-tube numbers collected from different flowers from the same 10 plants (based on Aizen [2005]). The curve represents the nonlinear regression of Eq. 1 for 1119 open-pollinated flowers collected during 1997 and 1998 (b ¼ 0.191, 95% CI ¼ 0.134–0.287; d ¼ 0.597, 95% CI ¼ 0.562–0.633). The gray dots represent the combined data for both years.
ers received primarily self-pollen, so that their relatively low fruit set in response to natural pollination largely reflects quality, rather than quantity, limitation. In particular, the relation of the observations (gray symbols) and the dose–response curve indicates that importing more pollen of the same quality would not have enhanced fruit set by most flowers. Instead, quality limitation reduced fruit set by 0.160 6 0.061 fruits per flower (mean 6 SE) compared to that following pollination with only high-quality cross-pollen. This interpretation of quality, but not quantity, limitation of fruit set is consistent with
the fact that Tristerix species mature only one seed per fruit, despite producing multiple ovules (Kuijt 1988). CONSEQUENCES OF POLLEN QUALITY FOR THE SUPPLEMENTATION TEST OF POLLEN LIMITATION Shortcomings of the standard test for pollen limitation associated with the effects of pollen quality can be illustrated with the assistance of Eq. 1. Suppose that during open pollination a stigma receives pO pollen grains per ovule, which includes a mixture of low-quality grains (perhaps self-pollen) and high-quality grains.
CONCEPTS & SYNTHESIS
276
MARCELO A. AIZEN AND LAWRENCE D. HARDER
Ecology, Vol. 88, No. 2
FIG. 3. Schematic relationship of seed production per flower to pollen delivery by pollen vectors for stigmas that are unmanipulated (dashed black curve: Eq. 1), or receive subsequent, supplemental cross-pollen (solid black curve: Eq. 4). The solid horizontal line (dXO) indicates maximal seed production following cross-pollination, and the dotted horizontal line (dOO) represents maximal seed production after open pollination. The gray circle illustrates seed production resulting from open pollination throughout a flower’s life, and the gray arrow depicts the response to subsequent supplemental pollination (gray square). Earlier pollen supplementation, when the flower had received less pollen naturally (open circle), would have resulted in a larger response (black square). Thus, comparison of open-pollinated and early-supplemented flowers will detect the difference indicated by the black arrow.
This pollen fertilizes ovules as described above, so that OebO pO ovules remain unfertilized. An experimenter then saturates the stigma with high-quality outcross pollen, denoted by subscript X, which fertilizes all these remaining ovules, so production of seeds sired by supplemental pollen depends on dX, but not bX. Note that any natural pollination occurring after supplementation cannot contribute to ovule fertilization or seed production. As a result, this pistil produces a total of S ¼ dO O 1 ebO pO þ dX O ebO pO ð4Þ ¼ dO O þ ðdX dO Þ ebO pO O seeds. As expected, supplemental cross-pollination elevates seed production, as long as some ovules remain unfertilized when supplemental pollen is added (compare solid and dashed curves in Fig. 3). However, contrary to the implicit assumption of the supplementalpollination technique, maximal seed production following supplementation varies negatively with the amount of pollen delivered by pollen vectors prior to supplementation (Fig. 3, solid curve). This decline arises because only ovules fertilized by high-quality supplemental pollen can increase seed production above dOO and the number of these ovules declines as prior open pollination increases. Fig. 3 illustrates two misleading features of pollen supplementation as a means of testing for pollen limitation when pollen quality affects seed production.
First, supplemental pollination will generally underestimate the combined limitations of pollen quality and quantity by (dX dO)(1 ebO pO )O seeds, as long as stigmas receive some pollen, as illustrated by the solid curve in Fig. 3 lying below dXO. Underestimation occurs because receipt of poor-quality pollen during open pollination usurps some ovules, precluding complete assessment of the maximal production of outcrossed seeds. Second, supplemental pollination overestimates the limitation of seed production associated with insufficient open pollination, as illustrated by the solid curve in Fig. 3 lying above dOO. This overestimate, (dX dO)(ebO pO )O, results because supplemental pollination confounds assessment of quantity limitation with some quality limitation. The overestimation of quantity limitation will be particularly severe if supplemental pollen is applied early during a flower’s life, before vectors have delivered much pollen. Obviously, this latter problem would be aggravated if pollen supplementation occurred before flowers had imported as much open pollen as possible during their lives as pollen recipients. These considerations reveal that pollen supplementation can assess quantity limitation accurately only if it adds pollen of the same quality as flowers receive naturally. Given that the quality of pollen delivered by vectors is usually unknown, plus the other problems described by Ashman et al. (2004), the pollensupplementation technique cannot be justified as a general method for either assessing the magnitude of
February 2007
POLLEN LIMITATION
277
PLATE 1. Reproductive stages during which pollen quality can contribute to pollen limitation, including (a) pollen germination and tube growth and (b) seed development. Panel (a) illustrates pollen tubes emerging from pollen grains on an Erythronium americanum stigma. Photo credit: B. M. Smith. Panel (b) shows unfertilized ovules, aborted seeds, and developing seeds in a Caragana arborescens maturing pod. Photo credit: L. D. Harder.
REPRESENTATIVE ASSESSMENT
OF
POLLEN LIMITATION
An appropriate approach to assessing total pollen limitation, including the contributions of quantity and quality limitation, is evident from Fig. 1b and exemplified in Fig. 2. To estimate both quantity limitation (kN: Eq. 2) and quality limitation (kQ: Eq. 3) maximal seed production must be measured following both open pollination (dLO ¼ dOO) and cross-pollination (dHO ¼ d X O), where the subscripts L and H stand for pollination with low- and high-quality pollen, respectively. The latter parameter is assessed easily by handpollinating flowers from which pollen vectors have been excluded with cross-pollen. In contrast, dOO must be estimated by fitting Eq. 1 to observations of the relation of seed production to pollen receipt for open-pollinated flowers (Fig. 1b, dotted curve). Ideally, this nonlinear regression analysis would use maximum-likelihood techniques to estimate b and d, which would recognize that seed production by a flower is a binomial process. Such an analysis would incorporate a binomial (rather than normal) error distribution, with the likelihood of observing S seeds in an ovary with O ovules given by O q S ð1 qÞOS ; S where q ¼ d(1 ebp ). Care should be taken in each of the preceding measurements of seed set to avoid the unwanted effects of preferential resource allocation to vigorous fruits within plants, which can compromise assessment of fecundity responses to pollination quantity and quality (see Ashman et al. 2004). Given
estimates of dOO and dXO, and the average seed production by open-pollinated flowers, SO, quantity limitation is estimated as kN ¼ dOO SO and quality limitation is estimated as kQ ¼ dXO dOO (note that neither estimate depends on b). Total pollen limitation represents the sum of kN and kQ. Thus, accurate assessment of the incidence of pollen limitation and the relative importance of quantity vs. quality limitation requires an integration of studies of pollination and post-pollination processes, which have generally been considered in isolation. The proposed technique may be difficult to apply in cases of extreme quantity limitation. When stigmas typically receive few pollen grains per ovule most observations will lie within the rapidly ascending portion of the dose–response curve rather than the asymptotic portion. However, pollen receipt varies extensively among plants within populations, even those subject to severe quantity limitation (e.g., Levin 1990, Herrera 2002), so that a large sample of flowers should usually include enough stigmas with numerous pollen grains to estimate asymptotic fecundity. CONCLUDING DISCUSSION The components of pollen limitation To illustrate the quantity and quality components of pollen limitation and their consequences we employed a saturating negative-exponential model of the relation of seed production to pollen receipt by individual flowers (Eq. 1). Like any model, this characterization incorporates various simplifying assumptions about the processes that it depicts, raising questions about the model’s general applicability. However, Fig. 1b reveals that the
CONCEPTS & SYNTHESIS
quantity limitation or differentiating quantity from quality limitation.
CONCEPTS & SYNTHESIS
278
MARCELO A. AIZEN AND LAWRENCE D. HARDER
essential features of our conception of quantity and quality limitation represent common aspects of seed production. Key among these common aspects are the expectations that: (1) seed production increases asymptotically with increasing pollen receipt (reviewed by Mitchell [1997], also see Aizen and Basilio 1998: Fig. 2); (2) quantity limitation (kN) represents a depression in seed production below the maximum (i.e., asymptote) expected for a given quality of pollen (Ashman et al. 2004, Knight et al. 2005); and (3) quality limitation (kQ) represents the difference in seed production with complete fertilization by low- vs. premium-quality pollen. Thus, most conclusions exposed by our model, including those concerning the efficacy of pollen supplementation as a test of pollen limitation, are not idiosyncratic consequences of the details of our model, but instead represent general features of pollen limitation. Our conception of quantity and quality limitation reflects one possible perspective. Pollen limitation depends on a population’s current state, including the species’ floral design and display, the size, density, and genetic structure of the population, and the composition and abundance of the pollinator fauna and co-flowering plant assemblage. Nevertheless, the occurrence and magnitude of pollen limitation varies among years for individual populations (Burd 1994, Ashman et al. 2004), primarily because of variation in pollinator service. Thus, from an ecological perspective pollen limitation varies temporally probably because of quantity limitation, which occurs against an underlying background of quality limitation. This perspective is represented by our depiction of quantity and quality limitation, as we propose that quantity limitation could be alleviated by receipt of more pollen of the same quality than pollinators currently deliver. This is also the perspective that has motivated the use of supplemental pollination to assess pollen limitation. In contrast, an evolutionary perspective would lead to a different partitioning of overall pollen limitation. For example, consider the outcome represented by the closed symbol in Fig. 1b. If this plant had a different floral design and display, genetic neighborhood, etc., so that it instead received the same number of high-quality pollen grains, then quality limitation could be conceived of as the difference in fecundity between the open and closed symbols. In this conception quantity limitation represents the extent to which the open symbol lies below the asymptotic fecundity for the highest quality pollen (dHO). This perspective may be useful in some analyses of the evolution of pollen limitation; however, we have emphasized the ecological perspective because it is more representative of the conception implicit in most studies of pollen limitation. Quantity and quality limitation, as we have defined them, constrain seed production additively (Fig. 1b). These components of pollen limitation depend on rather different aspects of the reproductive process. Quantity limitation incorporates effects during all phases of seed
Ecology, Vol. 88, No. 2
production, including ovule production (O), pollen import (P), fertilization (b), and seed development (d: Eq. 2), whereas quality limitation depends on only ovule production and seed development (Eq. 3). Note in particular, that pollinator limitation (low P) is only one component of quantitative pollen limitation, so that these terms cannot be used synonymously, although this practice is common (e.g., Bierzychudek 1981, Horvitz and Schemske 1988, A˚gren 1996). Fig. 1b also illustrates a second basic difference between these two components: quantity limitation is intense only at the lowest range of pollen receipt, when each arriving pollen grain has a high probability of fertilizing an ovule, whereas the severity of quality limitation remains unchanged throughout the range of pollen receipt. Thus, unlike quantity limitation, quality limitation is probably a chronic pollination problem faced by many plants, especially self-compatible species. The most obvious remedy to quantity limitation involves increased pollen import; however, this solution requires increased pollen transport in the population as a whole, and so necessitates improvements in both pollen export and import (Harder and Routley 2006). Pollen dispersal in animal-pollinated plants involves two components: visitation frequency (e.g., Aizen 2001, Kasagi and Kudo 2003) and the efficiency of individual pollinators in removing pollen and transporting it to conspecific stigmas (Wilson and Thomson 1991, but see Va´zquez et al. 2005). Thus, severe quantity limitation may select for enhanced floral signals and rewards, or changes in flower morphology that promote closer and more precise contact between pollinators’ bodies and anthers and stigmas. From the male side, quantitative pollen limitation may favor particular stamen architectures, and forms of pollen packaging and dispensing (Harder and Thomson 1989). The presence of b in Eq. 2 identifies aspects of fertilization success as additional essential components of quantity limitation. By determining the proportion of ovules that are fertilized, the processes involved—namely pollen grain germination, tube growth, gamete discharge, and syngamy—govern the number of zygotes with the potential to become seeds. At least some of these processes (e.g., pollen-tube growth) can have a strong genetic component (reviewed by Skogsmyr and Lankinen [2002]), so that insufficient pollen receipt could select for more permissive pollen–pistil interactions, perhaps explaining the repeated loss of self-incompatibility in many angiosperm lineages (e.g., Whisler and Snow 1992, Goodwillie 1999, 2001, Busch 2005). Furthermore, spatial and temporal variation in these post-pollination components, caused by either genetic differences among plants or a variety of environmental stresses (see Skogsmyr and Lankinen 2002), might influence the extent of quantitative pollen limitation among individuals, populations, or even species. Both quantity and quality limitation probably depend on the details of a plant’s mating environment, including
February 2007
POLLEN LIMITATION
experience less quality limitation, other than that caused by biparental inbreeding. Ovule production is the final influence on seed production incorporated in Eqs. 2 and 3. This term exposes the possibility of ovule limitation, which occurs if all ovules are fertilized, but too few embryos survive development to consume the available maternal resources (Harder and Routley 2006). Limitation of seed production by ovule production and pollen quality have identical consequences; however, ovule limitation arises because flowers produce too few ovules, whereas quality limitation results from poor performance of the existing ovules. Like quality limitation, ovule limitation is a poorly appreciated constraint on seed production, which is addressed in greater detail elsewhere (Harder and Routley 2006; L. D. Harder, M. B. Routley, and S. A. Richards, unpublished manuscript). The prevalence of quantity and quality limitation Recent reviews of pollen-supplementation experiments demonstrate that the pollination regimes of plants commonly do not allow them to achieve their reproductive potential (Burd 1994, Larson and Barrett 2000, Ashman et al. 2004, Knight et al. 2005). Indeed, pollen limitation may occur more often than these surveys indicate, as no response to pollen supplementation is expected when seed production is limited by pollen quality alone (Fig. 3). Furthermore, pollen supplementation overestimates the severity of quantity limitation if plants also experience quality limitation (Fig. 3). Thus, our model raises questions about the incidence and intensity of quantity limitation, rather than about whether pollen limitation occurs. Knight et al.’s (2005) meta-analysis of pollen-supplementation studies may illustrate the limited information that this technique provides about pollen limitation. They tested 11 hypotheses that considered various aspects of reproduction, primarily involving plant traits that could affect pollen import. Using phylogenetically independent contrasts, Knight et al. found no evidence for nine hypotheses and weak support (P ¼ 0.04) for the remaining two hypotheses, although sample sizes were limited for some tests. Given our interpretation of the outcome of pollen supplementation, a likely explanation for the inability of tests with adequate samples to detect significant effects is that pollen supplementation often does not measure any biologically relevant variable. Instead, pollen supplementation will typically measure unknown fractions of quantity and quality limitation (the length of the black arrow in Fig. 3), the magnitude of which will depend on specific circumstances. Thus, responses to pollen supplementation probably vary largely randomly among the studies surveyed by Knight et al. (2005), resulting in the limited support for their largely reasonable hypotheses. If this interpretation is correct, assessment of the influences of plant traits on interspecific variation in quantitative pollen limitation
CONCEPTS & SYNTHESIS
the abundance and efficiency of pollen vectors, and the abundance and dispersion of suitable mates. Different types of pollinators could have contrasting effects on quantity and quality limitation, depending on their delivery of self-pollen and/or pollen from related plants in genetically structured populations. For example, Brunet and Sweet (2006) found that hawk moths delivered a lower proportion of self-pollen than other pollinators of Aquilegia coerulea, because they were more likely to visit female-phase flowers before malephase flowers within an inflorescence, which should result in higher seed production for a given stigmatic pollen load. Herrera (1987, 2000) similarly found that Lepidoptera cause higher quality pollination of Lavandula latifolia than do bees or flies, because of their longer flights between plants. He also found that bees delivered more pollen, so that quantitative and qualitative aspects of pollination varied negatively. Such a negative relation between pollen quality and quantity may be a common feature of relying on a few pollinator species that deliver high-quality pollen, rather than a diverse pollinator fauna, because the beneficial pollinators of specialized plants represent a subset of possible pollinators. In addition to pollinator differences, quantity and quality limitation may vary with population size and density. For example, Wagenius (2006) found that seed production by Echinacea angustifolia declined with nearest-neighbor distance within populations and varied negatively among populations with population size. Reduced plant density commonly diminishes the rate at which pollinators visit plants, which should aggravate quantity limitation, and increases number of flowers visited by each pollinator, which should increase the incidence of self-pollination between flowers and aggravate quality limitation (see Grindeland et al. 2005). Thus, quantity and quality limitation probably vary extensively within and among populations because of local differences in the composition of the pollinator fauna and their responses to plant dispersion. According to Eqs. 2 and 3, factors influencing seed development can contribute to both quantity and quality limitation, although their effects differ. Quantity limitation depends on a plant’s absolute capacity to mature zygotes into seeds given the quality of the pollen that it actually receives (dL). In contrast, quality limitation is determined by the difference in the maximal production of seeds sired by pollen delivered by pollen vectors vs. the best-possible pollen (dH dL). Predispersal (early acting) inbreeding depression is probably the most important cause of seed-development failure in the context of quality pollen limitation (see Husband and Schemske 1996), whereas resource availability can also play a role in quality limitation. As a result, hermaphroditic plants with self-compatibility or late-acting (ovarian) self-incompatibility are vulnerable to both quantity and quality limitation, whereas dioecious species or those with self-incompatibility mechanisms that act in the stigma or style should
279
CONCEPTS & SYNTHESIS
280
MARCELO A. AIZEN AND LAWRENCE D. HARDER
awaits measurement of the true incidence of this cause of reproductive failure. The usual interpretation that the common enhancement of seed production by pollen supplementation indicates widespread quantity limitation is enigmatic, because chronic quantity limitation should select for traits that enhance pollen import (see Haig and Westoby 1988). Instead, we expect that the increased seed production commonly induced by pollen supplementation results largely from more successful seed development caused by supplementation with higher-quality pollen than flowers receive naturally. Despite being largely ignored from an ecological perspective (although see Ramsey 1995, Herrera 2000, Ramsey and Vaughton 2000, Anderson and Hill 2002, Finer and Morgan 2003, Ashman et al. 2004), pollen quality may often limit seed production. Self-pollination is probably the most widespread cause of quality limitation in self-compatible populations with high genetic loads. Goodwillie et al. (2005) found that self-fertilization accounted for 20–80% of the seeds produced by 42% of the 345 species for which information was available. They proposed that this high incidence of mixed mating often represents unavoidable self-pollination among flowers on multiflowered inflorescences and self-pollination that occurs after outcrossing to maximize ovule fertilization (reproductive assurance). Even self-incompatible species can suffer quality limitation as a result of inbreeding in a genetically structured population (e.g., Sun and Ritland 1998, Alves et al. 2003, Brennan et al. 2005). Given that partially and wholly outcrossing species commonly exhibit strong inbreeding depression during seed development (Husband and Schemske 1996), such selfpollination and/or biparental inbreeding must frequently depress seed production. These considerations emphasize that studies of pollen limitation would benefit from an expanded perspective that considers the ecological consequences of pollination quality. ACKNOWLEDGMENTS We thank N. P. Chacoff, J. W. Fox, and M. B. Routley for helpful discussion, and R. Aguilar, T.-L. Ashman, V. Chalcoff, C. M. Herrera, C. J. Melia´n, D. P. Va´zquez, N. M. Waser, and two anonymous reviewers for thoughtful comments on the manuscript. This research was completed while M. A. Aizen was a Sabbatical Fellow sponsored by the Fulbright Commission of Argentina at the National Center for Ecological Analysis and Synthesis, a Center funded by NSF (Grant number DEB-0072909), the University of California, and the Santa Barbara campus. Research funding was also provided by the Natural Sciences and Engineering Research Council of Canada (L. D. Harder). M. A. Aizen is a career researcher of the Consejo Nacional de Investigaciones Cientı´ ficas y Tecnolo´gicas of Argentina (CONICET). LITERATURE CITED A˚gren, J. 1996. Population size, pollinator limitation, and seed set in the self-incompatible herb Lythrum salicaria. Ecology 77:1779–1790. Aizen, M. A. 2001. Flower sex ratio, pollinator abundance, and the seasonal pollination dynamics of a protandrous plant. Ecology 82:127–144.
Ecology, Vol. 88, No. 2
Aizen, M. A. 2003. Influences of animal pollination and seed dispersal on winter flowering in a temperate mistletoe. Ecology 84:2613–2627. Aizen, M. A. 2005. Breeding system of Tristerix corymbosus (Loranthaceae), a winter-flowering mistletoe from the southern Andes. Australian Journal of Botany 53:357–361. Aizen, M. A., and A. Basilio. 1998. Sex differential nectar secretion in protandrous Alstroemeria aurea (Alstroemeriaceae): Is production altered by pollen removal and receipt? American Journal of Botany 85:245–252. Alves, R. M., A. S. Arter, A. M. Sebbenn, and F. Antonio. 2003. Mating system in a natural population of Theobroma grandiflorum (Willd. ex Spreng.) Schum., by microsatellite markers. Genetics and Molecular Biology 26:373–379. Anderson, G. J., and J. D. Hill. 2002. Many to flower, few to fruit: the reproductive biology of Hamamelis virginiana (Hamamelidaceae). American Journal of Botany 89:67–78. Ashman, T.-L., T. M. Knight, J. A. Steets, P. Amarasekare, M. Burd, D. R. Campbell, M. R. Dudash, M. O. Johnston, S. J. Mazer, R. J. Mitchell, M. T. Morgan, and W. G. Wilson. 2004. Pollen limitation of plant reproduction: ecological and evolutionary causes and consequences. Ecology 85:2408– 2421. Bierzychudek, P. 1981. Pollinator limitation of plant reproductive effort. American Naturalist 117:838–840. Brennan, A. C., S. A. Harris, and S. J. Hiscock. 2005. Modes and rates of selfing and associated inbreeding depression in the self-incompatible plant Senecio squalidus (Asteraceae): a successful colonizing species in the British Isles. New Phytologist 168:475–486. Brown, B. J., R. J. Mitchell, and S. A. Graham. 2002. Competition for pollination between an invasive species (purple loosestrife) and a native congener. Ecology 83:2328– 2336. Brunet, J., and H. R. Sweet. 2006. Impact of insect pollinator group and floral display size on outcrossing rate. Evolution 60:234–246. Burd, M. 1994. Bateman’s Principle and plant reproduction: the role of pollen limitation in fruit and seed set. Botanical Review 60:83–139. Busch, J. W. 2005. The evolution of self-incompatibility in geographically peripheral populations of Leavenworthia alabamica (Brassicaeae). American Journal of Botany 92: 1503–1512. Chacoff, N. P., and M. A. Aizen. 2006. Edge effects on flowervisiting insects in grapefruit plantations bordering premontane subtropical forest. Journal of Applied Ecology 43:18–27. Charlesworth, D., and B. Charlesworth. 1987. Inbreeding depression and its evolutionary consequences. Annual Review of Ecology and Systematics 18:237–268. Finer, M. S., and M. T. Morgan. 2003. Effects of natural rates of geitonogamy on fruit set in Asclepias speciosa (Apocynaceae): evidence favoring the plant’s dilemma. American Journal of Botany 90:1746–1750. Fre´ville, H., B. Colas, M. Riba, H. Caswell, A. Mignot, E. Imbert, and I. Olivieri. 2004. Spatial and temporal demographic variability in the endemic plant species Centaurea corymbosa (Asteraceae). Ecology 85:694–703. Goodwillie, C. 1999. Multiple origins of self-compatibility in Linanthus section Leptosiphon (Polemoniaceae): phylogenetic evidence from internal-transcribed-spacer sequence data. Evolution 53:1387–1395. Goodwillie, C. 2001. Pollen limitation and the evolution of selfincompatibility in Linanthus (Polemoniaceae). International Journal of Plant Sciences 162:1283–1292. Goodwillie, C., S. Kalisz, and C. G. Eckert. 2005. The evolutionary enigma of mixed mating systems in plants: occurrence, theoretical expectation and empirical evidence. Annual Review of Ecology, Evolution, and Systematics 36: 47–79.
February 2007
POLLEN LIMITATION
Larson, B. M. H., and S. C. H. Barrett. 2000. A comparative analysis of pollen limitation in flowering plants. Biological Journal of the Linnean Society 69:503–520. Levin, D. A. 1990. Sizes of natural microgametophyte populations in pistils of Phlox drummondii. American Journal of Botany 77:356–363. Lloyd, D. G., and D. J. Schoen. 1992. Self- and crossfertilization in plants. I. Functional dimensions. International Journal of Plant Sciences 153:358–369. Mitchell, R. J. 1997. Effects of pollination intensity on Lesquerella fendleri seed set: variation among plants. Oecologia 109:382–388. Morgan, M. T., W. G. Wilson, and T. M. Knight. 2005. Plant population dynamics, pollinator foraging, and the selection of self-fertilization. American Naturalist 166:169–183. Murphy, S. D. 2000. Field testing for pollen allelopathy: a review. Journal of Chemical Ecology 26:2155–2172. Obeso, J. R. 2004. A hierarchical perspective in allocation to reproduction from whole plant to fruit and seed level. Perspectives in Plant Ecology, Evolution and Systematics 6: 217–225. Parker, I. M. 1997. Pollinator limitation of Cytisus scoparius (Scotch broom), an invasive exotic shrub. Ecology 78:1457– 1470. Ramsey, M. 1995. Ovule preemption and pollen limitation in a self-fertile perennial herb (Blandfordia grandiflora, Liliaceae). Oecologia 103:101–108. Ramsey, M., and G. Vaughton. 2000. Pollen quality limits seed set in Burchardia umbellata (Colchicaceae). American Journal of Botany 87:845–852. Sage, T. L., R. I. Bertin, and E. G. Williams. 1994. Ovarian and other late-acting self-incompatability systems. Pages 116–140 in E. G. Williams, A. E. Clarke, and R. B. Knox, editors. Genetic control of self-incompatability and reproductive development in flowering plants. Kluwer Academic Publishers, Dordrecht, The Netherlands. Silvertown, J., M. Franco, I. Pisanty, and A. Mendoza. 1993. Comparative plant demography—relative importance of lifecycle components to the finite rate of increase in woody and herbaceous perennials. Journal of Ecology 81:465–476. Skogsmyr, I., and A˚. Lankinen. 2002. Sexual selection: an evolutionary force in plants? Biological Reviews 77:537– 562. Sun, M., and K. Ritland. 1998. Mating system of yellow starthistle (Centaurea solstitialis), a successful colonizer in North America. Heredity 80:225–232. Va´zquez, D. P., W. Morris, and P. Jordano. 2005. Interaction frequency as a surrogate for the total effect of animal mutualists on plants. Ecology Letters 8:1088–1094. Wagenius, S. 2006. Scale dependence of reproductive failure in fragmented Echinacea populations. Ecology 87:931–941. Waser, N. M., and M. V. Price. 1991. Outcrossing distance effects in Delphinium nelsonii: pollen loads, pollen tubes, and seed set. Ecology 72:171–179. Whisler, S. L., and A. A. Snow. 1992. Potential for the loss of self-incompatibility in pollen-limited populations of mayapple (Podophyllum peltatum). American Journal of Botany 79: 1273–1278. Wilson, P. S., and J. D. Thomson. 1991. Heterogeneity among floral visitors leads to discordance between removal and deposition of pollen. Ecology 72:1503–1507.
CONCEPTS & SYNTHESIS
Grant, V. 1981. Plant speciation. Columbia University Press, New York, New York, USA. Griffin, C. A. M., and C. G. Eckert. 2003. Experimental analysis of biparental inbreeding in a self-fertilizing plant. Evolution 57:1513–1519. Grindeland, J. M., N. Sletvold, and R. A. Ims. 2005. Effects of floral display size and plant density on pollinator visitation rate in a natural population of Digitalis purpurea. Functional Ecology 19:383–390. Haig, D., and M. Westoby. 1988. On limits to seed production. American Naturalist 131:757–759. Harder, L. D., and S. C. H. Barrett. 1996. Pollen dispersal and mating patterns in animal-pollinated plants. Pages 140–190 in D. G. Lloyd and S. C. H. Barrett, editors. Floral biology: studies on floral evolution in animal-pollinated plants. Chapman and Hall, New York, New York, USA. Harder, L. D., and M. B. Routley. 2006. Pollen and ovule fates and reproductive performance by flowering plants. Pages 61– 80 in L. D. Harder and S. C. H. Barrett, editors. The ecology and evolution of flowers. Oxford University Press. Harder, L. D., and J. D. Thomson. 1989. Evolutionary options for maximizing pollen dispersal of animal-pollinated plants. American Naturalist 133:323–344. Herlihy, C. R., and C. G. Eckert. 2004. Experimental dissection of inbreeding and its adaptive significance in a flowering plant, Aquilegia canadensis (Ranunculaceae). Evolution 58: 2693–2703. Herrera, C. M. 1987. Components of pollinator ‘‘quality’’: comparative analysis of a diverse insect assemblage. Oikos 50:79–90. Herrera, C. M. 2000. Flower-to-seedling consequences of different pollination regimes in an insect-pollinated shrub. Ecology 81:15–29. Herrera, C. M. 2002. Censusing natural microgametophyte populations: variable spatial mosaics and extreme finegraininess in winter-flowering Helleborus foetidus (Ranunculaceae). American Journal of Botany 89:1570–1578. Heywood, J. S. 1991. Spatial analysis of genetic variation in plant populations. Annual Review of Ecology and Systematics 22:335–355. Horvitz, C. C., and D. W. Schemske. 1988. A test of the pollinator limitation hypothesis for a neotropical herb. Ecology 69:200–206. Husband, B. C., and D. W. Schemske. 1996. Evolution of the magnitude and timing of inbreeding depression in plants. Evolution 50:54–70. Kasagi, T., and G. Kudo. 2003. Variations in bumble bee preference and pollen limitation among neighboring populations: comparisons between Phyllodoce caerulea and Phyllodoce aleutica (Ericaceae) along snowmelt gradients. American Journal of Botany 90:1321–1327. Kawagoe, T., and N. Suzuki. 2005. Self-pollen on a stigma interferes with outcrossed seed production in a self-incompatible monoecious plant, Akebia quinata (Lardizabalaceae). Functional Ecology 19:49–54. Knight, T. M., J. A. Steets, J. C. Vamosi, S. J. Mazer, M. Burd, D. R. Campbell, M. R. Dudash, M. O. Johnston, R. J. Mitchell, and T.-L. Ashman. 2005. Pollen limitation of plant reproduction: pattern and process. Annual Review of Ecology, Evolution, and Systematics 36:467–497. Kuijt, J. 1988. Revision of Tristerix (Loranthaceae). Systematic Botany Monographs [American Society of Plant Taxonomists], Volume 19.
281
