Search

Comparing traits of invasive species within and beyond their ancestral range may improve our understanding of processes that promote aggressive spread. Solanum elaeagnifolium (silverleaf nightshade) is a noxious weed in its ancestral range in North America and is invasive on other continents. We compared investment in flowers and ovules, pollination success, and fruit and seed set in populations from Arizona, USA (“AZ”) and Greece (“GR”). In both countries, the populations we sampled varied in size and types of present-day disturbance. Stature of plants increased with population size in AZ samples whereas GR plants were uniformly tall. Taller plants produced more flowers, and GR plants produced more flowers for a given stature and allocated more ovules per flower. Similar functional groups of native bees pollinated in AZ and GR populations, but visits to flowers decreased with population size and we observed no visits in the largest GR populations. As a result, plants in large GR populations were pollen-limited, and estimates of fecundity were lower on average in GR populations despite the larger allocation to flowers and ovules. These differences between plants in our AZ and GR populations suggest promising directions for further study. It would be useful to sample S. elaeagnifolium in Mediterranean climates within the ancestral range (e.g., in California, USA), to study asexual spread via rhizomes, and to use common gardens and genetic studies to explore the basis of variation in allocation patterns and of relationships between visitation and fruit set.

Pollination success of animal-pollinated flowers depends on rate of pollinator visits and on pollen deposition per visit, both of which should vary with the pollen and nectar "neighborhoods" of a plant, i.e., with pollen and nectar availability in nearby plants. One determinant of these neighborhoods is per-flower production of pollen and nectar, which is likely to respond to environmental influences. In this study, we explored environmental effects on pollen and nectar production and on pollination success in order to follow up a surprising result from a previous study: flowers of Ipomopsis aggregata received less pollen in years of high visitation by their hummingbird pollinators. A new analysis of the earlier data indicated that high bird visitation corresponded to drought years. We hypothesized that drought might contribute to the enigmatic prior result if it decreases both nectar and pollen production: in dry years, low nectar availability could cause hummingbirds to visit flowers at a higher rate, and low pollen availability could cause them to deposit less pollen per visit. A greenhouse experiment demonstrated that drought does reduce both pollen and nectar production by I. aggregata flowers. This result was corroborated across 6 yr of variable precipitation and soil moisture in four unmanipulated field populations. In addition, experimental removal of pollen from flowers reduced the pollen received by nearby flowers. We conclude that there is much to learn about how abiotic and biotic environmental drivers jointly affect pollen and nectar production and availability, and how this contributes to pollen and nectar neighborhoods and thus influences pollination success.

In western North America, hummingbirds can be observed systematically visiting flowers that lack the typical reddish color, tubular morphology, and dilute nectar of "hummingbird flowers." Curious about this behavior, we asked whether these atypical flowers are energetically profitable for hummingbirds. Our field measurements of nectar content and hummingbird foraging speeds, taken over four decades at multiple localities, show that atypical flowers can be as profitable as typical ones and suggest that the profit can support 24-h metabolic requirements of the birds. Thus, atypical flowers may contribute to successful migration of hummingbirds, enhance their population densities, and allow them to occupy areas seemingly depauperate in suitable resources. These results illustrate what can be gained by attending to the unexpected.

“Pollination syndromes” are specific combinations of floral traits that are proposed to evolve convergently across angiosperm lineages in response to different types of animal pollinators. In spite of their long history, pollination syndromes have not been tested adequately–they rarely have been examined critically to determine how well they describe floral trait diversity or predict pollinators. In a recent meta-analysis of data from the literature, Rosas-Guerrero et al. (2014) provide a welcome test that draws on insights from past studies. At the same time, their study illustrates several difficulties of meta-analysis approaches in general, and for pollination biology in particular. Here we discuss those difficulties and propose some solutions. We first consider how to gather studies from the literature without introducing unintended bias, such as the old-fashioned method of working backward from cited literature. We next consider how to deal with difficulties that invariably arise when extracting and analyzing often-incomplete information from heterogeneous studies. Finally we discuss issues of interpreting and presenting the results in the most informative manner. We conclude that although Rosas-Guerrero et al. (2014) and other studies such as Ollerton et al. (2009) have arrived at different conclusions about the utility of pollination syndromes, their results are not necessarily incompatible.

Individual differences in fecundity often serve as proxies for differences in overall fitness, especially when it is difficult to track the fate of an individual's offspring to reproductive maturity. Using fecundity may be biased, however, if density-dependent interactions between siblings affect survival and reproduction of offspring from high- and low-fecundity parents differently. To test for such density-dependent effects in plants, we sowed seeds of the wildflower Ipomopsis aggregata (scarlet gilia) to mimic partially overlapping seed shadows of pairs of plants, one of which produced twice as many seeds. We tested for differences in offspring success using a genetic marker to track offspring to flowering multiple years later. Without density dependence, the high-fecundity parent should produce twice as many surviving offspring. We also developed a model that considered the geometry of seed shadows and assumed limited survivors so that the number of juvenile recruits is proportional to the area. Rather than a ratio of 2∶1 offspring success from high- versus low-fecundity parents, our model predicted a ratio of 1.42∶1, which would translate into weaker selection. Empirical ratios of juvenile offspring and of flowers produced conformed well to the model's prediction. Extending the model shows how spatial relationships of parents and seed dispersal patterns modify inferences about relative fitness based solely on fecundity.

Premise of research. Dust particles and pollen grains are similar in size. Dust deposition might therefore influence the pollination and reproduction of flowering plants. Little is known about such effects, however, despite more general interest in ecological effects of dust.Methodology. We used observational and experimental methods to explore whether dust generated by traffic on unpaved roads affects the amounts of pollen received and numbers of seeds produced by four species of native wildflowers in the western United States.Pivotal results. Flowers of Nuttall’s larkspur (Delphinium nuttallianum), scarlet gilia (Ipomopsis aggregata), Lewis flax (Linum lewisii), and sulphur paintbrush (Castilleja sulphurea) growing 1–2 m from a road received substantially more dust and less pollen than those growing 40–50 m away. We observed the same pattern when we transplanted individuals of the first two species into pots and placed pots near to compared with far from a road. Experimental “hand dusting” of scarlet gi...

High-elevation ecosystems are expected to be particularly sensitive to climate warming because cold temperatures constrain biological processes. Deeper understanding of the consequences of climate change will come from studies that consider not only the direct effects of temperature on individual species, but also the indirect effects of altered species interactions. Here we show that 20 years of experimental warming has changed the species composition of graminoid (grass and sedge) assemblages in a subalpine meadow of the Rocky Mountains, USA, by increasing the frequency of sedges and reducing the frequency of grasses. Because sedges typically have weak interactions with mycorrhizal fungi relative to grasses, lowered abundances of arbuscular mycorrhizal (AM) fungi or other root-inhabiting fungi could underlie warming-induced shifts in plant species composition. However, warming increased root colonization by AM fungi for two grass species, possibly because AM fungi can enhance plant water uptake when soils are dried by experimental warming. Warming had no effect on AM fungal colonization of three other graminoids. Increased AM fungal colonization of the dominant shrub Artemisia tridentata provided further grounds for rejecting the hypothesis that reduced AM fungi caused the shift from grasses to sedges. Non-AM fungi (including dark septate endophytes) also showed general increases with warming. Our results demonstrate that lumping grasses and sedges when characterizing plant community responses can mask significant shifts in the responses of primary producers, and their symbiotic fungi, to climate change.

Spatial gradients in human activity, coyote activity, deer activity, and deer herbivory provide an unusual type of evidence for a trophic cascade. Activity of coyotes, which eat young mule deer (fawns), decreased with proximity to a remote biological field station, indicating that these predators avoided an area of high human activity. In contrast, activity of adult female deer (does) and intensity of herbivory on palatable plant species both increased with proximity to the station and were positively correlated with each other. The gradient in deer activity was not explained by availabilities of preferred habitats or plant species because these did not vary with distance from the station. Does spent less time feeding when they encountered coyote urine next to a feed block, indicating that increased vigilance may contribute, along with avoidance of areas with coyotes, to lower herbivory away from the station. Judging from two palatable wildflower species whose seed crop and seedling recruitment were greatly reduced near the field station, the coyote-deer-wildflower trophic cascade has the potential to influence plant community composition. Our study illustrates the value of a case-history approach, in which different forms of ecological data about a single system are used to develop conceptual models of complex ecological phenomena. Such an iterative model-building process is a common, but underappreciated, way of understanding how ecological systems work.

The link between reproductive and vegetative ecology of flowering plants is rarely explored, despite its importance for understanding population processes and fitness. This link can be studied by using experimental or natural variation in seed input to the soil to assess how reproductive success affects vital rates of offspring. We previously reported for Ipomopsis aggregata that per-seed probability of germinating is insensitive to density of seeds sown into plots, whereas per capita flower production among adults that grow from the seedlings declines in nonlinear fashion with density. Here we describe a parallel non-experimental study. We related seedling emergence to estimated natural seed input (‘seed rain’) in three populations across ten summers and monitored seedlings that emerged in the first two summers throughout their life histories. Seedling emergence in 1996 was linearly related to seed rain from plants that flowered in 1995. This density independent seed-to-seedling transition recurred over the next nine summers, but the slope varied with springtime precipitation. Total numbers of 1996 seedlings that survived to flower and numbers of flowers they produced increased linearly with seed rain in one population, but did not vary detectably in the other two, consistent with negative density dependence. In consequence λ (the dominant eigenvalue of a population projection matrix) decreased from high values at low densities of seed rain to a relatively constant low value with greater seed rain. We also detected density dependence in the 1995 seedling cohort in survival and flower production. The similarity of results from natural and experimental studies supports a conclusion of nonlinear density dependence and shows that characterizing it requires the full life history. For this plant species and others, studies of pollination and fecundity alone may not suffice to draw conclusions about population change or fitness.

Anthropogenic climate change is widely expected to drive species extinct by hampering individual survival and reproduction, by reducing the amount and accessibility of suitable habitat, or by eliminating other organisms that are essential to the species in question. Less well appreciated is the likelihood that climate change will directly disrupt or eliminate mutually beneficial (mutualistic) ecological interactions between species even before extinctions occur. We explored the potential disruption of a ubiquitous mutualistic interaction of terrestrial habitats, that between plants and their animal pollinators, via climate change. We used a highly resolved empirical network of interactions between 1420 pollinator and 429 plant species to simulate consequences of the phenological shifts that can be expected with a doubling of atmospheric CO(2). Depending on model assumptions, phenological shifts reduced the floral resources available to 17-50% of all pollinator species, causing as much as half of the ancestral activity period of the animals to fall at times when no food plants were available. Reduced overlap between plants and pollinators also decreased diet breadth of the pollinators. The predicted result of these disruptions is the extinction of pollinators, plants and their crucial interactions.

Although herbivory can occur throughout a plant’s life, little is known about relative fitness impacts of damage at different life stages. In the long-lived monocarpic wildflower, Ipomopsis aggregata (scarlet gilia), for example, the response to browsing by ungulates in the year of flowering has been studied extensively, whereas damage and its fitness consequences during the preceding years of vegetative growth remain largely unexplored. As part of a long-term demographic study of I. aggregata , we mapped 5324 individual seedlings belonging to two annual cohorts in three natural populations, and followed these plants throughout their lives. Of the 30.4% of plants that survived past seedling stage, 15.3% suffered observable damage to the apical meristem of their single rosette of leaves, usually resulting in multiple rosettes in vegetative plants and multiple flowering stalks in plants that survived to flower. Vegetative damage reduced survival to flowering by ca 30%, and slowed growth, leading to an average delay in flowering of over one yr relative to undamaged plants. On average, damaged plants were 136% larger than undamaged members of their cohort in the summer before flowering, but produced only 86% as many flowers, and achieved 78% the overall fitness compared to undamaged plants when damage effects were integrated across their life history. Previously published studies from the same sites suggested that plants suffering browsing during only the year of flowering, achieved 47% of the fitness of unbrowsed plants. Our results indicate that plants are better able to recover from vegetative damage than from loss of their initial reproductive effort, perhaps because of the longer time over which recovery is possible in the former case. These results underscore the importance of studying the consequences of damage across the entire life history.

Pollination by animals is critical to sexual reproduction of most angiosperms. However, little is known about variation in pollination service to single plant species. We report results of a long-term study of Ipomopsis aggregata, a semelparous montane herb whose flowers are visited by hummingbird and insect pollinators as well as “floral larcenists.” We censused flower visitors over seven summers at permanent study sites separated by several hundred meters, and counted pollen delivered to flowers on a subset of plants observed for visitation. The species composition of the community of visitors varied significantly across years and within the flowering season; sites varied significantly only in the magnitude of parallel annual changes in the visitor community. Rates of flower visitation fluctuated over an order of magnitude or more. Variation in mean stigma pollen load among plants flowering in the same site and year was explained by a causal path model in which visitation rates by pollinators and larcen...

Changes in the distribution and abundance of bird and small mammal species at urban-wildland edges can be caused by different factors. Edges can affect populations directly if animals respond behaviorally to the edge itself or if proximity to edge directly affects demographic vital rates (an “ecotonal” effect). Alternatively, urban edges can indirectly affect populations if edges alter the characteristics of the adjacent wildland vegetation, which in turn prompts a response to the altered habitat (a “matrix” or “habitat” effect). We studied edge effects of birds and small mammals in southern Californian coastal sage scrub, and assessed whether edge effects were attributable to direct behavioral responses to edges or to animal responses to changes in habitat at edges. Vegetation species composition and structure varied with distance from edge, but the differences varied among study sites. Because vegetation characteristics were correlated with distance from edge, responses to habitat were explored by using independently-derived models of habitat associations to calibrate vegetation measurements to the habitat affinities of each animal species. Of sixteen species examined, five bird and one small mammal species responded to edge independently of habitat features, and thus habitat restoration at edges is expected to be an ineffective conservation measure for these species. Two additional species of birds and one small mammal responded to habitat gradients that coincided with distance from edge, such that the effect of edge on these species was expressed via potentially reversible habitat degradation.

Depletion of experimental seed patches by granivorous animals often is used as a qualitative assay of foraging activity. An optimal foraging model suggests that seed amounts remaining when foragers leave patches (“giving-up-density”, GUD) also provide quantitative measures of foraging economics, diet strategies and foraging abilities. Such quantitative uses of GUDs rest on several largely untested assumptions. We tested two of these with Merriam’s kangaroo rats: that gain curves are smoothly decelerating, and that foragers leave patches at a constant harvest rate. Harvest rates indeed declined with patch residence time, but in the piecewise linear fashion expected of systematic search. Animals also revisited areas within patches less frequently than expected with random search. In the field, they depleted patches in multiple visits and did not use a constant-rate leaving rule. These deviations from model assumptions cast doubt on inferences about foraging ecology that have been based on quantitative GUD theory.

Distributions of related species along environmental gradients provide ecologists with insights into factors that limit distributions of species. We apply this approach to two species of kangaroo rats. A survey in 1983 showed that Dipodomys merriami is replaced by D. panamintinus at the tops of elevational gradients in parts of the Mojave Desert where both species occur. Over the 17 y since the initial survey we have conducted six additional censuses along one such gradient. In years of high population densities D. merriami is more abundant at low and D. panamintinus at high elevations along this gradient. Following periods of drought, however, when population densities are reduced overall, D. merriami expands upward along the gradient, whereas D. panamintinus becomes restricted to even higher elevations. This pattern suggests that D. merriami normally is restricted to lower elevations by competition from the larger D. panamintinus, and experiences competitive release when the latter is at low de...

Individual species may be useful as indicators of biodiversity if an association exists between the presence of a species and another component of biodiversity. We evaluated 40 species of birds and small mammals, including 11 species of conservation concern, as potential indicators of species richness and spe- cies composition in southern California coastal sage scrub habitats. This habitat, which is the target of large- scale conservation planning, has been greatly reduced by human development and supports many plants and animals of conservation concern. We asked whether there is an association between the presence of a po- tential indicator species and the species richness and composition of the bird or small-mammal community in which it is found. We used point counts and live-trapping to quantify the distribution of birds and small mammals, respectively, at 155 points in 16 sites located in three counties. Of the few species we found associ- ated with species richness, some were associated with higher species richness and others with lower richness, and species of conservation concern were not more frequently associated with species richness than were common species. Ordination analysis revealed a geographic gradient in coastal sage scrub bird and small- mammal species composition across southern California, and 18 of the species we evaluated were associated with the composition of the bird and small-mammal community in which they were found. Our results sug- gest that efforts to conserve bird and small-mammal biodiversity in coastal sage scrub should not focus exclu- sively on rare species or on locations with the highest species richness, but instead should focus on a diverse suite of species that are representative of the range of variation in communities found in coastal sage scrub habitats.

Diversity of species in communities of heteromyid rodents presents a classic problem to ecologists, because species are similar ecologically and share a limiting seed resource. Mechanisms of coexistence considered to date have focused on interspecific variation in ability to exploit heterogeneity in resources caused by environmental factors. An unexplored possibility is that coexistence is promoted by heterogeneity among species in seed-caching behavior. To begin evaluating this possibility, we asked whether coexisting species differ in their propensity to cache and in types of caches made. In an indoor arena, we presented millet seeds to 8 species of kangaroo rats (Dipodomys) and pocket mice (Perognathus and Chaetodipus) from 2 communities, 1 in California and 1 in Arizona. Species within communities differed in amounts of seed consumed and cached per night. Both consumption and caching increased with body mass in a manner similar to whole-animal metabolic rate, suggesting that energetics underl...

In their response to Ollerton et al.’s (2015) cautions about methods used by Rosas-Guerrero et al. (2014) to test the pollination syndromes, Aguilar et al. (2015) criticize an earlier paper by Ollerton et al. (2009). Here we respond to their concerns.

Increasing “greenhouse” gases are predicted to warm the earth by several degrees Celsius during the coming century. At high elevations one likely result is a longer snow-free season, which will affect plant growth and reproduction. We studied flowering and fruiting of 10 angiosperm species in a subalpine meadow over 4 yr, focusing on plant responses to warming by overhead heaters. The 10 species reproduced in a predictable sequence during 3–4 mo between spring snowmelt and fall frosts. Experimental warming advanced the date of snowmelt by almost 1 wk on average, relative to controls, and similarly advanced the mean timing of plant reproduction. This phenological shift was entirely explained by earlier snowmelt in the case of six plant species that flowered early in the season, whereas four later-flowering species apparently responded to other cues. Experimental warming had no detectable effect on the duration of flowering and fruiting, even though natural conditions of early snowmelt were associated with ...

Biology, like all sciences, is increasingly compartmentalized. This helps us to interact more easily with fellow specialists, but also tends to shield us from useful cross-fertilization with other fields. For example, plant and animal ecologists have established largely distinct research traditions over the years. Botanists studying population ecology, interspecific competition, and ecological and evolutionary aspects of sexual reproduction can benefit by importing more of the conceptual advances occurring in zoology (and vice versa). Botanists working on plant-animal interactions have a similar opportunity to shed light on their questions by learning more about the animals. For example, botanists studying pollination may be unaware of some modern advances in animal physiology and behaviour which undermine a typological view of pollination systems. Similarly, botanists studying loss of seeds to granivores may be unaware of details of animal behaviour that can cause granivores to benefit the plants in surprising ways. These examples illustrate how improved communication with zoology can enrich plant ecology. To remove barriers to communication we suggest some individual and collective actions that plant ecologists can take.

We examined habitat use by the endangered Stephens’ kangaroo rat ( Dipodomys stephensi) in different seasons and assessed whether this may be influenced by a sympatric congener, the Pacific kangaroo rat (Dipodomys agilis). Trapping on three plots over 2 years revealed these species were rarely captured at the same trap stations. Spatial segregation was highly significant when both species were at high density. The spatial distribution of these species was temporally stable where both species were relatively abundant, but where D. agilis was relatively uncommon the distribution of D. stephensi varied from one census to another. The abundance of three microhabitats (grass, debris, and bare ground) followed a regular seasonal pattern of variation that was consistent across 2 years of substantially different rainfall. A canonical discriminant analysis showed that the five quantified microhabitats (those above and bush and rock cover) provided highly significant discrimination between the trap stations occupied by the two species. Dipodomys stephensi was associated with trap stations where grass cover and bare ground were abundant but where bush and rock were uncommon. Dipodomys agilis was associated with stations that had large amounts of bare ground and average abundances of bush and rock cover. The spatial segregation of these species appears to be mediated by habitat preferences; D. stephensi prefers grassland and D. agilis prefers sage scrub. This suggests that habitat management for D. stephensi should include (1) controlling the spread of shrubs into grassland and (2) creating dispersal corridors of open habitat to link areas of suitable habitat where none presently exist. Each of these options may be needed to maintain viable populations in all reserves designated for the conservation of D. stephensi. Examinamos el uso del habitat por la rata canguro de Stephen en peligro de extincion ( Di-podomys stephensi) en diferentes estaciones y evaluamos si esto podria estar influenciado por un congenere simpatrico, la rata canguro del Pacifico (Dipodomys agilis). Trampeos en tres areas a lo largo de dos anos revelaron que estas especies fueron raramente capturadas en las mismas estaciones de muestreo. La segregacion espacial fue altamente significativa cuando ambas especies presentaron alta densidad. La distribucion espacial de las especies fue temporalmente estable cuando ambas especies fueron relativamente abundantes, pero donde D. agilis fue relativamente poco comun, la distribucion de D. stephensi vario de un censo a otro. La abundancia de tres microhabitats ( pasto, detritus y suelo desnudo) siguieron un patron estacional de variacion regular que fue consistente por dos anos con precipitacion pluvial sustancialmente diferente. Un analisis de descriminacion canonica mostro que los cinco habitats cuantificados (los antes mencionados mas cobertura de arbusto y de roca) proporcionaron alta descriminacion significativa entre las estaciones de trampeo ocupadas por las dos especies. D. stephensi estuvo asociada con estaciones donde el pasto y suelo desnudo fueron abundantes, pero donde arbustos y rocas fueron poco comunes. D. agilis estuvo asociada con areas que presentaron grandes cantidades de suelo desnudo y abundancia promedio de arbustos y rocas. La segregacion espacial de estas especies parece ser mediada por sus preferencias de habitat; D. stephensi prefiere pastizales y D. agilis prefiere matorrales de salvia. Esto sugiere que el manejo de habitat para D. stephensi debe incluir (1) control de la dispersion de arbustos en pastizales y (2) crear corredores de dispersion de habitat abierto para comunicar ares con habitat viable donde no existen actualmente. Cada una de estas opciones podria ser necesaria para mantener poblaciones viables en todas las reservas designadas para la conservacion de D. stephensi.

Patterns of resource availability mold many ecological processes, but we know little about the availability of resources to consumers in nature, even for well-studied systems such as the granivorous animals of North American deserts. What we do know about seed resources in deserts is based primarily on seeds extracted from soil samples, but this might present a distorted view of resource availability if animals mostly harvest newly produced seeds before they enter the soil seed bank. In order to assess how large the distortion might be, we simultaneously monitored the seed bank and “seed rain” over a 19-mo period in the eastern Mojave Desert of California. The seed bank averaged ≈106000 seeds/m2 and 38 g/m2, much higher than values reported for other North American desert sites. This corresponds roughly to the seed production of a single year, since daily seed rain averaged 262 seeds/m2 and 0.26 g/m2. However, input from the seed rain did not accumulate in the soil. Instead, the seed bank decreased by a daily average of 114 seeds/m2 and 0.007 g/m2 during our study. This suggests that virtually all seeds germinate, die, or are harvested by granivores soon after being dispersed. Large seeds comprised a greater fraction of the seed rain than of the seed bank, suggesting that such seeds are differentially depleted, probably by granivores, before they enter the soil. Because seed drop was seasonal, temporal variation comprised a significant component of among-sample variance in the seed rain. Temporal variance in the seed bank was much smaller, presumably because granivores harvested most of the seed rain. Conversely, spatial variance was a significant component for the seed bank, but not the seed rain, perhaps as a result of spatial patterns of seed harvest or seed caching by granivores. By virtue of these variance patterns, as well as other attributes, seeds in the soil present different challenges to granivores than do newly produced seeds. Our understanding of desert granivore foraging and community ecology, and of granivore–seed interactions, depends critically on choosing the appropriate measure of seed availability to granivores.

The form of angiosperm flowers is thought to have evolved in part via selection that excludes ineffective pollinators and increases the efficiency of pollen transfer by effective ones. In previous studies with the montane plant Ipomopsis aggregata, we documented pollinator-mediated selection on several aspects of floral form, including flower width. This character varies continuously within natural populations of I. aggregata, and individuals with wider flowers export more pollen per flower to surrounding plants. We previously showed that this component of phenotypic selection is due primarily to the per- visit effectiveness of hummingbirds, the most important pollinator in our study populations. Here we investigate mechanisms of differential visit effectiveness in greater detail. First, counts of pollen in unvisited flowers showed that pollen production itself increases with width. In aviary experiments, hummingbirds also removed a greater proportion of available pollen as width increased. We next videotaped visits to flowers that varied in width either naturally, or by experimental treatment, and found that hummingbirds inserted their bills more deeply into wider flowers. Finally, we directly manipulated how deeply birds could insert their bills, and found that more pollen was removed after deep insertion. Thus, several mechanisms appear to underlie selection on corolla width via visit effectiveness in pollen export. One involves a phenotypic correlation with pollen production; this underscores the value of experiments for untangling indirect from direct selection. Another mechanism involves direct selection due to the depth of insertion of a hummingbird's bill; this is rare evidence for one form of selection involving the "fit" between pollinator and flower.

One view of pollination systems is that they tend toward specialization. This view is implicit in many discussions of angiosperm evolution and plant-pollinator coevolution and in the long-standing concept of pollination syndromes. But actual pollination systems often are more generalized and dynamic than these traditions might suggest. To illustrate the range of specialization and generalization in pollinators' use of plants and vice versa, we draw on studies of two floras in the United States, and of members of several plant families and solitary bee genera. We also summarize a recent study of one local flora which suggests that, although the colors of flowers are aggregated in phenotype space, there is no strong association with pollinator types as pollination syndromes would predict. That moderate to substantial generalization often occurs is not surprising on theoretical

Events that follow pollination, such as pollen-tube growth and seed maturation, comprise an important phase of angiosperm reproduction. Differential success during this "postpollination" phase may represent phenotypic selection, including sexual selection, or interaction between parents caused, for example, by their genetic similarity. By providing a detailed partitioning of variance in success, diallel crossing designs offer great potential to determine which processes are occurring and their relative magnitudes. We performed three partial diallels with the montane herb Ipomopsis aggregata, using a large sample of parental plants (69 total). Embedded in the designs were crossing-distance treatments of 1 m, 10 m, and 100 m, reflecting a range of parental genetic similarity. We partitioned phenotypic variance in seed set per fruit into six components using restricted maximum-likelihood (REML) analysis. For one diallel, we also partitioned variance in seed mass into five components, and estimated two components of covariance between seed set and mass. Variance caused by maternal effects (Vmat ) comprised 12%-35% of total variance in seed set and 62% of variance in seed mass, and there was a significant negative environmental covariance between seed set and seed mass. Parental interaction made no detectable contribution to phenotypic variance in either of our measures of postpollination success, although crossing distance did contribute slightly but significantly to fit of the model in some cases. Finally, there was no detectable paternal variance (Vpat ) in seed set or seed mass. These results are in keeping with reports from other studies of natural plant populations. The finding of little or no paternal variance in particular suggests little scope for postpollination sexual selection through the male function of cosexual plants such as I. aggregata.

Movements by Dipodomys stephensi were assessed by live-trapping and radiotelemetry to evaluate consequences of habitat fragmentation for this endangered species. The median of the maximum distances moved between captures was 29.2 m for the 557 individuals with ≥2 captures. The median distance between first and last monthly home-range centers was 17.6 m for individuals captured in ≥2 months. Males moved longer distances than did females; lactating females were especially sedentary. Dispersal distances were similar for juveniles and adults and for two sites with different geometries of habitat patches. The median distance between first and last home-range center of adults radiotracked for 15–127 days was 9.5 m, indicating stable home ranges. However, radiotracked individuals moved greater maximum distances than most live-trapped individuals apparently moved over a lifetime. Much dispersal, therefore, occurs within an area about which animals have knowledge. This should facilitate habitat selection and decrease the cost of dispersal in terms of fitness. The frequency distribution of maximum dispersal distances underestimated the frequency of moves >400 m. Two processes may have contributed to this bias. First, the frequency of long-distance moves was underestimated because we measured dispersal away from a single grid. Second, our trapping grids were located in homogeneous habitat without dirt roads, which appear to facilitate long-distance movement. Unvegetated corridors show promise for increasing the connectedness of local populations of D. stephensi . Connectedness should be considered in management of endangered species such as D. stephensi , because the expected persistence of metapopulations is sensitive to the extent of dispersal and the fitness of dispersers.

Effective conservation of endangered species often is hampered by inadequate knowledge of demography. We extracted information on survival and fecundity from an 18-month, live-trapping study of Dipodomys stephensi, and from this we developed an age-structured demographic model to assess population viability. Adult Stephens’ kangaroo rats persisted longer than juveniles, and adult females persisted longer than adult males. Disappearance rates were high in the first months after initial capture. Thereafter, the fraction of animals persisting decreased slowly and in an approximately linear fashion on a semilogarithmic scale, suggesting age-independent mortality factors such as predation. Juvenile persistence did not differ substantially between two years of strikingly different rainfall. Onset of breeding followed the start of winter rains. Length of the breeding season, average number of litters per female, and the fraction of first-year females breeding were much greater in the year of higher rainfall. We propose a birth-pulse demographic model for D. stephensi that distinguishes juvenile and adult age classes. Temporal environmental variation can be modeled adequately with a constant survivorship schedule and variable fecundity determined by yearly precipitation. Several issues should be resolved, however, before conservation decisions are based on the model. Better estimates of juvenile survivorship are critical, the quantitative relationship between precipitation and fecundity must be determined, and the potential for density dependence and source-sink population dynamics must be evaluated.

Depending on its genetic causes, outbreeding depression in quantitative characters may occur first in the free-living F1 generation produced by a wide cross. In 1981-1985, we generated F1 progenies by hand-pollinating larkspurs (Delphinium nelsonii) with pollen from 1-m, 3-m, 10-m, or 30-m distances. From the spatial genetic structure indicated by previous electrophoretic and reciprocal transplantation studies, we estimate that these crosses range from being inbred (f ≈ 0.06) to outbred. We planted 594 seeds from 66 maternal sibships under natural conditions. As of 1992, there was strong evidence for both inbreeding depression and outbreeding depression. Progeny from intermediate crossing distances grew approximately twice as large as more inbred or outbred progeny in the first 5 yr after planting (P = 0.013, repeated measures ANOVA), and survived almost 1 yr longer on average (contrast of 3-m and 10-m treatments versus 1 m and 30 m; P = 0.028, ANOVA). Twenty maternal sibships produced flowering individuals; only four and two of these represented 1-m and 30-m crossing distances, respectively (P = 0.021, G-test). The cumulative fitness of intermediate distance sibships averaged about twice that of 1-m sibships, and five to eight times that of 30-m sibships (P = 0.017, ANOVA). Thus, even though progeny of 1-m crosses were inbred to a degree only about one-eighth that of selling, inbreeding depression approximated 50%, and outbreeding depression equaled or exceeded 50% for all fitness components.

The shrub Encelia farinosa (Asteraceae) exhibits geographic variation in aboveground architecture and leaf traits in parallel with environmental variation in temperature and moisture. Measurements of plants occurring across a natural gradient demonstrated that plants in desert populations produce smaller, more pubescent leaves and are more compact and branched than plants in more mesic coastal environments. This phenotypic variation is interpreted in part as adaptive genetic differentiation; small size and pubescence reduce leaf temperature and thus increase water-use efficiency but at the cost of lower photosynthetic rate, which results in slower growth and more compact growth form. We explored the basis of phenotypic variation by planting seed offspring from coastal and desert populations in common gardens in both environments. Phenotypic differences among populations persisted in both common gardens, suggesting a genetic basis for trait variation. Desert offspring outperformed coastal offspring in the desert garden, suggesting superior adaptation to hot, dry conditions. Herbivore damage was greater for all offspring in the coastal garden. Phenotypic characters also showed plastic responses; all offspring had smaller, more pubescent leaves and more compact growth form in the desert garden. Our results confirm that leaf size and pubescence are heritable characters associated with pronounced variation in plant architecture.

Based on previous studies, extreme (>99%) self-sterility in scarlet gilia (Ipomopsis aggregata) appears to be involved in late-acting ovarian self-incompatibility (OSI). Here, we confirm this suggestion by comparing structural events that follow from cross- vs. self-pollinations of I. aggregata. Growth of cross- and self-pollen tubes in the style at 11 h and growth in the ovary at 24 h was equivalent. Nonetheless, by 24 h, cross-pollen effected a significantly higher percentage of both ovule penetration and fertilization. Ovules in self-pollinated flowers showed pronounced changes, including an absence of embryo sac expansion and reduced starch in the integument, by 11 h post-pollination, well before pollen tube entry into the ovary. In addition, the integumentary tapetum and adjacent 1-3 cell layers exhibited abnormal cell division, pronounced deposition of thick, pectin-rich cell walls, and cellular collapse. Ovules and embryo sacs from cross-pollinated flowers rarely showed such features. Developmental changes in ovules from self-pollinated flowers eventually resulted in integument and embryo sac collapse, a process not observed in ovules of unpollinated flowers. We suggest that OSI involves long-distance signaling between self-pollen or self-pollen tubes and carpel tissue that reduces availability of receptive ovules for fertilization before pollen tubes arrive in the ovary.

The rodent family Heteromyidae contains bipedal hoppers and quadrupedal runners. The possibility that bipedalism is associated with forelimb specialization for nonlocomotory functions, such as burrowing and seed-gathering, motivated a static functional-morphometric and interspecific allometric analysis of 18 metric characters of the forelimb skeleton. A principal-components analysis, across 28 species in six genera, showed that lengths of proximal (scapula, humerus) and distal (ulna, radius, metacarpal) elements were negatively allometric, and widths were positively allometric. Quadrupedal and bipedal species groups showed qualitatively similar allometric patterns, except that scapula width anterior to the spine was positively allometric in quadrupeds and negatively allometric in bipeds; scapula width posterior to the spine was positively allometric in bipeds and isometric in quadrupeds; and olecranon length was isometric in bipeds and positively allometric in quadrupeds. Most morphometric characters varied significantly among species within genera, even when effects of size variation were reduced by reconstructing all species to a common general size (as indicated by their score on the first principal component). These shape differences caused species to vary in the mechanical advantage of the forelimb, of possible importance for digging and seed-harvesting performance. Relative to quadrupeds, bipedal species tended to have greater mechanical advantage for proximal forelimb elements and smaller mechanical advantage for distal forelimb elements, but only the distal pattern remained in reconstructed forms, and no functional character was significantly different when tested over variation among genera nested within locomotion type. Cluster analysis confirmed that forelimb characters related to digging or seed-harvest are not coincident with mode of locomotion. Forelimb characters were, however, associated with digging or seed-harvest performance. Mechanical advantage of the proximal forelimb was positively related to an index of the compaction of soils with which 26 desert-dwelling species are associated, and also to relative use of heavy vs. light soils by nine species in the laboratory. Across 10 species, deviations in seed-harvest rate from expected allometric values were negatively correlated with mechanical advantage of the distal forelimb.

Tropical dry deciduous forest is an endangered ecosystem whose plant-pollinator relationships are little known. We characterised a portion of the web of interactions between flowering plants and flower visitors in the Kirindy Forest of the Menabe region of west-central Madagascar. Taking a plant-centered approach, we observed individuals of the 5 most abundant native plant species that were coming into flower at the end of the annual dry season, and recorded all identifiable flower-visitors. Taking a visitor-centered approach, we walked a network of established trails and listened for distinctive calls of a common flower-visiting bird, noting the plant species visited. The former approach revealed connections among the early-flowering species via birds and insects, whereas the latter confirmed these connections and added an additional plant species. Flowers of the 6 plant species were visited on average by 5.5 animal species, while 10 visitor species for which we had reasonable samples frequented on average the flowers of 3.3 plant species. These qualitative results resemble those reported from other temperate and tropical webs, in that interactions appeared to be relatively generalised by pollinator species and body plan (e.g., birds vs. bees). Also in agreement, the visitation web was significantly nested, with more-specialised species tending to interact with mutualistic partners that were themselves more generalised. In addition to documenting previously-unreported interactions, therefore, this preliminary web conforms to more widespread patterns emerging for pollination systems at the community level.

North American heteromyid rodents are alike in their primary seed harvesting behavior (scratch—digging), which is a potentially important component of their foraging energetics. The present study was designed to quantify the energy cost and efficiency ([Benefit — Cost]/Cost) of scratch—digging throughout the body size range represented among the Heteroyidae. The resting energy expenditure of bipeds was generally greater than that of quadrupeds on an equal body size basis, whereas their energy expenditure during scratch—digging (DMR) was equivalent. Seed concentration and substrate bulk—density did not markedly affect DMR, but seed concentration did alter the rate of seed harvest. Species differences in seed harvest rate were correlated with mode of locomotion, supporting the hypothesis that bipedality is correlated with specialization of the forelimbs for seed handling. In heteromyids, resting metabolic rate (RMR) was elevated by a factor of 1.5 over basal metabolic rate (BMR); DMR was elevated by a factor of 3.4 over BMR and by a factor of 2.3 over RMR. DMR, seed harvest rate, and the energy cost of scratch—digging on a per seed basis scaled with body mass to the 0.65, 0.31, and 0.33 powers, respectively. Although large bipedal heteromyids harvested seeds at more than twice the rates of small quadrupedal heteromyids, the unequal scaling of DMR and harvest rate makes them less efficient at scratch—digging. In comparison with other heteromyids, the kangaroo mouse (Microdipodops megacephalus) was exceptionally efficient at scratch—digging. Its higher efficiency may allow it to forage on sparsely distributed seeds in open areas where the energy return would be inadequate for larger bipeds. The pocket mouse (Chaetodipus baileyi) was exceptionally inefficient at scratch—digging. Its low efficiency in sand may be correlated with specialization for foraging in pebbly soils. For heteromyids of a given size, factors affecting pouching rate are most likely to determine foraging efficiency because DMR is essentially a fixed cost.

Coexisting heteromyid rodent species of North American deserts differ in habitat use and in locomotory morphology. Quadrupedal species forage primarily in struc- turally complex microhabitats, such as under bush canopies, while bipedal species forage in open spaces. A common explanation for this morphology-microhabitat association is that species differing in morphology also differ in vulnerability to predators, that micro- habitat structure affects predation risk, and that animals preferentially forage in the safest microhabitats. We tested this for two bipedal and two quadrupedal heteromyid species (matched by body size), and one cricetid species, by quantifying effects of habitat and illumination on activity and on risk of predation by Great Homed Owls. Capture frequencies were lower for all heteromyid species than for the cricetid species, Peromyscus maniculatus. Heteromyid activity was lower in open habitat and under bright illumination. Illumination had no significant effect on risk, perhaps because rodents changed activity patterns under full moon to compensate for a potential increment in risk. Habitat, however, did affect risk: all species were attacked and captured more frequently in the open. Bipedal species were attacked relatively more in the open than were quadrupeds. If these results apply to all predators, they indicate that predation alone cannot account for the divergent microhabitat associations of bipedal and quadrupedal species. Bipedal het- eromyids, however, escaped owl attacks more frequently than did quadrupeds of equivalent size. It is therefore conceivable that they experience lower overall risk in nature, where owls may preferentially attack more easily captured prey species when given a choice. Under these circumstances, owl predation could reinforce divergent microhabitat special- izations based on some other factor, such as foraging economics, by restricting quadrupeds more strongly than bipeds to the safety of bushes.

In the hummingbird-pollinated herb Ipomopsis aggregata, selection through male function during pollination favors wide corolla tubes. We explored the mechanisms behind this selection, using phenotypic selection analysis to compare effects of corolla width on two components of male pollination success, pollinator visit rate and pollen exported per visit. During single visits by captive hummingbirds, flowers with wider corollas exported more pollen, and more dye used as a pollen analogue, to stigmas of recipient flowers. Corolla width was less strongly related to visit rate in the field, and had no direct effect on visit rate after nectar production and corolla length were controlled for. Moreover, the phenotypic selection differential was 80% higher for the effect on pollen exported per visit, suggesting that this is the more important mechanism of selection.

Seed production of a flower depends both on the quantity of pollen received and on its "quality" as manifested in rates of germination, ovule fertilization, and seed maturation. Disentangling effects of these two factors is complicated by the fact that seed set is a decelerating function of pollen quantity. One potential element of quality is the genetic similarity of pollen donor and recipient. In ten experiments with Delphinium nelsonii, seed set varied with outcrossing distance, a correlate of genetic similarity. Over all experiments, a 10-m distance significantly (P < .02) outperformed shorter and longer distances by 23-33%, suggesting quality differences. To explore this possibility, we fit decelerating negative-exponential regressions to relation- ships among stigma pollen load, pollen tube number reaching the ovary, and seed set in a subset of experiments having 1-, 10-, and 100-m treatments. Regression parameters sug- gested that 10-m pollen is most efficient at producing tubes and seeds, and analysis of residuals from pooled regressions (a nonlinear ANCOVA) showed that 1 0-m residuals were most positive. The latter effect was significant for pollen load-seed set relationships (P = .009, five experimental replicates) and pollen tube-seed set relationships (P = .021, two replicates), but not pollen load-pollen tube relationships (P = .128, three replicates). These results illustrate the utility of nonlinear regression in distinguishing pollen quan- tity and quality. They reinforce the conclusion that outcrossing distance affects pollen quality in D. nelsonii, and that an "optimal outcrossing distance" between 1 and 100 m maximizes seed set on average in pollinations using one donor per carpel. Insofar as quality differences persist in natural pollen mixtures that arrive on stigmas, outcrossing distance should affect reproductive success through both paternal and maternal sexual functions.

Despite extensive study of pollination and plant reproduction on the one hand, and of plant demography on the other, we know remarkably little about links between seed production in successive generations, and hence about long-term population consequences of variation in pollination success. We bridged this "generation gap" in Ipomopsis aggregata, a long-lived semelparous wildflower that is pollinator limited, by adding varying densities of seeds to natural populations and following resulting plants through their entire life histories. To determine whether pollen limitation of seed production constrains rate of population growth in this species, we sowed seeds into replicated plots at a density that mimics typical pollination success and spacing of flowering plants in nature, and at twice that density to mimic full pollination. Per capita offspring survival, flower production, and contribution to population increase (lambda) did not decline with sowing density in this experiment, suggesting that typical I. aggregata populations freed from pollen limitation will grow over the short term. In a second experiment we addressed whether density dependence would eventually erase the growth benefits of full pollination, by sowing a 10-fold range of seed densities that falls within extremes estimated for the natural "seed rain" that reaches the soil surface. Per capita survival to flowering and age at flowering were again unaffected by sowing density, but offspring size, per capita flower production, and lambda declined with density. Such density dependence complicates efforts to predict population dynamics over the longer term, because it changes components of the life history (in this case fecundity) as a population grows. A complete understanding of how constraints on seed production affect long-term population growth will hinge on following offspring fates at least through flowering of the first offspring generation, and doing so for a realistic range of population densities.

A biologically explicit simulation model of resource competition between two species of seed-eating heteromyid rodent indicates that stable coexistence is possible on a homogeneous resource if harvested food is stored and consumers steal each other's caches. Here we explore the coexistence mechanisms involved by analyzing how consumer phenotypes and presence of a noncaching consumer affect the competitive outcome. Without cache exchange, the winning consumer is better at harvesting seeds and produces more offspring per gram of stored food. With cache exchange, coexistence is promoted by interspecific trade-offs between harvest ability, metabolic efficiency, and ability to pilfer defended caches of heterospecifics or scavenge undefended caches of dead conspecifics or heterospecifics. Cache exchange via pilferage can equalize competitor fitnesses but has little stabilizing effect and leads to stable coexistence only in the presence of a noncaching consumer. In contrast, scavenging is both equalizing and stabilizing and promotes coexistence without a third consumer. Because body size affects a heteromyid rodent's metabolic rate, seed harvest rate, caching strategy, and ability to steal caches, interspecific differences in body size should produce the trade-offs necessary for coexistence. The observation that coexisting heteromyids differ in body size therefore indicates that cache exchange may promote diversity in heteromyid communities.

Ungulate browsing of flowering stalks of the semelparous herb Ipomopsis aggregata leads to regrowth of lateral inflorescences, a response that has been reported to yield overcompensation in some cases (browsed plants with higher reproductive success than unbrowsed), but undercompensation in others. Little is known about the mechanisms that cause such variable tolerance to herbivory. We explored one possible mechanism--variation in effects of browsing on pollination--by clipping I. aggregata inflorescences to mimic browsing, observing subsequent visits by pollinators and nectar-robbers, and adding pollen by hand to flowers of some clipped and unclipped plants. Clipping reduced floral display size and increased inflorescence branching, but neither hummingbirds, the primary pollinators, nor nectar-robbing bumblebees showed any preference for unclipped versus clipped plants. Clipping delayed flowering; this shift in phenology caused clipped plants to miss the peak of hummingbird activity and to have lower per-flower visitation rates than unclipped controls in one year, but to have greater overlap with birds and higher visitation rates in the subsequent year. In three sites and 2 years, clipped plants exposed to natural pollination suffered extreme undercompensation, producing on average only 16% as many seeds as unclipped controls. This was not directly attributable to clipping effects on pollination, however, because clipped plants were unable to increase fecundity when provided with supplemental pollen by hand. Taken altogether, our results suggest that compensation was constrained less by indirect effects of browsing on pollination than by its direct impacts on resource availability and hence on the ability of plants to regrow lost inflorescence tissue and to fill seeds. Exploring the physiological and developmental processes involved in regrowth of inflorescences and provisioning of seeds is a promising future direction for research designed to understand variation in browsing tolerance.

Outbreeding depression in progeny fitness may arise from disruption of local adaptation, disruption of allelic coadaptation, or a combination of these "environmental" and "physiological" mechanisms. Thus the minimum spatial scale over which outbreeding depression arises should depend on the spatial scale of gene dispersal and (with an environmental mechanism) of change in selection regimes. We previously reported substantial outbreeding depression in lifetime fitness of progeny resulting from crosses among parents separated by 100 m in natural populations of the herbaceous plant Ipomopsis aggregata. In this paper we explore the effect of crossing distance on fitness in two additional experiments begun in 1987 and 1990. We planted seed progeny derived from partial diallel crossing designs in randomized blocks in maternal environments and scored emergence of seedlings, survival, and eventual flowering of individuals over the subsequent six to eight years. Nested within each diallel design were crossing distances of 1 m, 10 m, and 100 m. Compared to 1-m and 10-m progeny, 100-m progeny of the 1987 diallel suffered a significant reduction in seedling emergence, and both 1-m and 100-m progeny that survived to flower achieved lower lambda-values on average than 10-m progeny. Total outbreeding depression suffered by 100-m relative to 10-m progeny was approximately 10%, compared to approximately 30% in our earlier study of I. aggregata. Progeny of 10-m crosses also outperformed 1-m and 100-m progeny of the 1990 diallel by approximately 5%, but no difference among crossing distance treatments was significant. Thus, the magnitude of outbreeding depression in 100-m crosses varied among experiments. This is not surprising given likely spatial and temporal variation in gene flow and selection regimes, different population histories, and different parental and progeny environments. Characterizing outbreeding depression on the shortest spatial scales over which it is expressed, as well as its variation and causes, is worthwhile because it promises to shed light on the earliest stages of angiosperm speciation.

Experimental manipulation of a trait can be used to distinguish direct selection from selection of correlated traits and to identify mechanisms of selection. Here we use experiments to investigate phenotypic selection of stigma position in angiosperm flowers. In natural populations of the subalpine herb Ipomopsis aggregata, plants with more strongly exserted stigmas receive more pollen per flower, indicating selection favoring stigma exsertion during the pollination stage of the life cycle. We pose four hypotheses for this association, two involving direct selection on stigma position and two involving indirect selection of a correlated floral trait. The first three hypotheses were tested using hand pollinations that mimicked natural hummingbird visitation, and by presenting captive hummingbirds with a series of flowers that differed in stigma and anther positions, sex ratio, and presence of anthers. In these experiments, pollen deposition either was independent of stigma exsertion or was highest on inserted stigmas, suggesting direct selection against exserted stigmas. In natural populations, however, stigma exsertion is highly correlated with time spent by the protandrous flowers in the pistillate phase. When we manipulated the latter trait in the field, pollen deposition increased with duration of exposure to hummingbirds, indicating indirect selection for stigma exsertion. Stigma exsertion and time spent in the pistillate phase are genetically and phenotypically correlated, as shown by a quantitative genetic experiment conducted in the field with paternal half sibships. Our results suggest that the evolution of stigma position can be driven by selection of a genetically correlated trait.

We used powdered fluorescent dyes to estimate receipt of self vs. outcross pollen in the self-incompatible species Ipomopsis aggregata (Polemoniaceae). Flowers on small and large plants received equal amounts of outcross pollen, whereas flowers on large plants received more self pollen, so the proportion of self pollen delivered through geitonogamy increased with plant size. In natural populations emasculation of all flowers on a plant raised average seed set per flower from 5.19 to 6.99 and also raised fruit set, though not significantly. From these results one expects a negative correlation between plant size and seeds per flower. The opposite trend was observed in a sample of plants in the field, suggesting that deleterious effects of geitonogamy on female fecundity in large plants can be overruled by other factors such as size-related fruit or seed abortion. Results are discussed in relation to the evolution of gynodioecy.