Is increased male flower production a strategy for avoidance of predispersal seed predation in andromonoecious plants?

Floral gender in angiosperms often varies within and among populations. We conducted a field survey to test how predispersal seed predation affects sex allocation in an andromonoecious alpine herb Peucedanum multivittatum. We compared plant size, male and perfect flower production, fruit set, and seed predation rate over three years among nine populations inhabiting diverse snowmelt conditions in alpine meadows. Flowering period of individual populations varied from mid‐July to late August reflecting the snowmelt time. Although perfect flower and fruit productions increased with plant size, size dependency of male flower production was less clear. The number of male flowers was larger in the early‐flowering populations, while the number of perfect flowers increased in the late‐flowering populations. Thus, male‐biased sex allocation was common in the early‐flowering populations. Fruit‐set rates varied among populations and between years, irrespective of flowering period. Fruit‐set success of individual plants increased with perfect flower number, but independent of male flower number. Seed predation by lepidopteran larvae was intense in the early‐flowering populations, whereas predation damage was absent in the late‐flowering populations, reflecting the extent of phenological matching between flowering time of host plants and oviposition period of predator moths. Seed predation rate was independent of male and perfect flower numbers of individual plants. Thus, seed predation is a stochastic event in each population. There was a clear correlation between the proportion of male flowers and the intensity of seed predation among populations. These results suggest that male‐biased sex allocation could be a strategy to reduce seed predation damage but maintain the effort as a pollen donor under intensive seed predation.     

Does pre-dispersal seed predation limit reproduction and population growth in the alpine clonal plant Geum reptans?

We studied the impact of the seed damaging gall midge larva Geomyia alpina on its perennial alpine host plant Geum reptans. We analysed the effect of seed predation on reproduction by seeds, i.e. seed number, seed mass, and seed viability and on growth and clonal propagation of non-protected plants in comparison to plants protected from predation by an insecticide. Additionally, we assessed the consequences of seed predation for population growth using matrix projection modelling. Seed predation resulted in a decrease in total seed mass per flower head by 23.8% in non-protected plants (P < 0.05). Individual seed mass decreased with increasing infestation intensity (P < 0.05). Seed number remained unaffected because the sucking feeding behaviour by gall midge larvae does not evoke seed abortion. Percent germination of seeds from non-protected plants was reduced by 97.9% compared to seeds from protected plants. According to reduced seed viability, modelling revealed a decrease in population growth rate from λ = 1.055 to λ = 1.041. Predation did neither influence total plant biomass nor biomass fractions. But stolon dry-weight of non-protected plants increased by 24.1% (P < 0.05), which may indicate a trade-off between sexual reproduction and clonal propagation. Our results demonstrate that despite substantial reduction of viable seeds, predation by gall midge larvae only slightly affected population growth of G. reptans suggesting that in this alpine species, persistence by longevity and clonal propagation can balance potential seed losses by predation, at least for local population growth.     

Flowering phenology of <Emphasis Type="Italic">Ulex europaeus</Emphasis>: ecological consequences of variation within and among populations

Reproductive phenology of gorse (Ulex europaeus L., Genisteae, Fabaceae) is unusual in that the onset and duration of flowering vary greatly among individuals within populations: some plants initiate flowering in autumn or winter and continue flowering through spring, others initiate flowering in early spring. To understand the origin of this diversity and its ecological consequences, we investigated flowering phenology of randomly sampled individuals from five different natural populations in Brittany (France). Reproductive success was evaluated for individuals with contrasting flowering patterns, from 16 natural populations. Flower production, pod production, seed production and seed predation were estimated. Plants initiating flowering in spring produced larger numbers of flowers and pods over a shorter period than plants flowering from winter to spring, which produced few flowers and pods at a time but over a longer period. Pod production of long-flowering plants did not differ significantly between winter and spring, but their pods were more intensively attacked by seed predators in spring than in winter. We discuss our results in relation to biotic and abiotic parameters. We postulate that long-flowering can be interpreted as a bet-hedging strategy, spreading the risk of pod failure (rotting or freezing) in winter and of seed predation in spring.     

Pre-dispersal seed predation in gynodioecious Geranium sylvaticum is not affected by plant gender or flowering phenology

Sex-specific interactions with antagonists may explain female maintenance in gynodioecious populations if seeds produced by hermaphroditic plants are preferred over seeds produced by female plants. Among antagonistic interactions, pre-dispersal seed predators have received relatively little attention even though they may exert sex-specific selective pressures on the evolution of floral and flowering traits. In this work, I investigate temporal variation in seed predation in gynodioecious Geranium sylvaticum, where in addition to female and hermaphrodite individuals, plants with an intermediate sexual expression are also present in most populations. Specifically, I examined whether seed predation is linked to flowering phenology, plant gender, and sexual dimorphism in floral and seed traits over the flowering season using an experimental field population. Within the population, I selected female, intermediate, and hermaphrodite plants with different timing of flowering onset (early, mid, or late), and collected seeds across the fruiting period. Seeds were weighed and examined for seed predator damage. The results show that the three genders experienced similar levels of seed predation attack regardless of their flowering phenology, and that overall seed predation was not related to changes in seed production or seed mass. These results suggest that sexual dimorphism in seed predation cannot be responsible for female maintenance in this species.     

Sex‐biased seed predation in gynodioecious Dianthus superbus var. longicalycinus (Capryophyllaceae) and differential influence of two seed predator species on the floral traits

Gynodioecy, the co‐occurrence of hermaphrodite and female individuals within a species, is maintained by differential reproductive success between sexes. Recently, researchers recognized that not only pollinators but also herbivores are important agents in the evolution and maintenance of gynodioecy, when herbivory is hermaphrodite biased. In this study, we investigated whether there is hermaphrodite‐biased herbivory in a gynodioecious plant, Dianthus superbus var. longicalycinus, and if so, what floral traits influenced hermaphrodite‐biased herbivory. We measured flower morphology (flower diameter, calyx tube length, corolla height and petal width) and phenology of flowers of female individuals, hermaphrodites and gynomonoecious individuals in a natural population. We also investigated seed predation and predator species. At the study site, Sibinia weevils (Curculionidae; Coleoptera) and Coleophora moths (Coleophoridae; Lepidoptera) were common pre‐dispersal seed predators in this species. The weevil appeared early in the flowering season, and weevil predation correlated with flower phenology. Because female individuals did not flower early in the season, weevil predation was less frequent in female individuals. Moth predation correlated with calyx length. The calyx length of flowers of female individuals was smaller than those of hermaphrodites, but a direct comparison of moth predation rates failed to find a significant difference among sex morphs. We found that the two seed predators had different effects on floral traits in D. superbus var. longicalycinus. We suggest that weevil predation contributes to the maintenance of gynodioecy because female individuals successfully escaped weevil predation by flowering late. It remains unclear why flower phenology is different among sex morphs.     

Recovery of aSasa tsuboiana population after mass flowering and death

The recovery process of aSasa tsuboiana population after a mass flowering and death in 1977 was investigated by 15 years of observation in the Hira Mountains, Kinki district, western Japan. Seed production was high (6600–13 800 seeds m⁻² inSasa plots and 3900 seeds m⁻² in a forest plot) but emergent seedling density was low (14–21 seedlings m⁻²), probably because of seed predation byMicrotus montebelli occurring between seed shedding and the next spring. The seedling density had decreased further by the next year and theS. tsuboiana population recovered from only a limited number of seedlings. In spite of such a low initial density, theS. tsuboiana population was able to regenerate successfully and attained the previous full stand height in 7–16 years.Miscantbus sinensis invaded and delayed the recovery ofS. tsuboiana in one plot, butS. tsuboiana became dominant as it caught up with the height ofM. sinensis. Seedling growth patterns, such as frequent tillering, the onset of rhizome extension in the early stage of seedling growth and frequent culm production from rhizomes, played important roles in the successful regeneration ofS. tsuboiana.     

Relationships between flowering phenology,plant size and reproductive success in shape Lotus corniculatus (Fabaceae)

Over three years the flowering phenology of individuals of Lotus corniculatus has been studied in relation to fruit set and seed predation to determine the relationships between four components of flowering time, plant size and reproductive success. Timings of first and peak flowering, and duration and synchrony of flowering differed between individuals in the same years. Between years, timing of first flowering was highly correlated for the same individuals, and was closely correlated with plant size and duration of flowering–larger plants flowered earlier and for a longer period. Peak flowering and synchrony were not correlated between-years for individuals.Fruit production and seed predation were correlated with some of the components of flowering phenology in some years, but not in others. The inconstancy of these relationships suggests that directional or stabilising selection is not acting consistently on the aspects of reproductive success studied in this work. The inconstancy of selection may result in the rather asynchronous flowering phenologies of individuals of L. corniculatus observed.We emphasize the importance of studying different components of flowering phenology in relation to individual plant size over several seasons. This work has shown that plant size not only has a direct effect on individual plant fecundity but also can influence flowering time and hence indirectly affect reproductive output.     

Temporal Variation in Seed Predation by Insects in a Population of Syagrus romanzoffiana (Arecaceae) in Santa Catarina Island, SC, Brazil

Insect seed predation may vary depending on seed production. The present study considers the hypothesis that the rates of seed predation tend to be smaller in years of higher fruit production. Thus, we monitored the production of fruits and predation of seeds of the palm Syagrus romanzoffiana over 2?years in the Atlantic Forest (Parque Municipal da Lagoa do Peri, Florianópolis, SC, Brazil), between July 2006 and June 2008. Plots of 0.25?m² were fitted under 20 mother plants and fruits were monthly collected for assessment of abundance and seed predation. There was variation in fruit production between the 2?years and among reproductive plants. Predation rates were high and occurred in the predispersal phase by the Curculionidae Revena rubiginosa Boheman, Anchylorhynchus aegrotus Fahraeus, and Anchylorhynchus variabilis Gyllenhal. Seed predation by these species of Anchylorhynchus is first registered in the present study. In average, about 60% of the seeds monthly produced in the population tend to escape insect predation in year of high or low production, becoming available for recruitment. The predation rate was not related to the amount of fruits produced per reproductive plant. Also, different than expected, there was a positive relation between the rates of seed predation and the total of fruits produced monthly on the plots. Thus, no evidence for the satiation of insect seed predators was found in this study with S. romanzoffiana.     

Temporal consistency in the spatial pattern of seed predation across a forest – old field edge

Seed predation is an important factor in determining the rate of tree establishment in abandoned agricultural land. Edges, through altered habitat use by small mammals, may influence the spatial pattern of forest regeneration in these successional sites. To determine the spatial pattern of seed predation across a forest-old field edge, we used a grid that began 30 m inside the forest and extended 60 m into the old field. Seed stations were placed at regular 10-m intervals and were monitored for removal of Acer rubrum seed for 50 d. This design was repeated over four years (1995–1998). Small mammal trapping was conducted in the final year of the study to determine the spatial pattern of seed predators within the site. Removal rates varied among the four years of the study with years of high and low removal rates. However, the spatial pattern of seed removal rate was similar in all years. Final survival ranged from 0.7–15.5% of seeds, with lower final survival in years with faster rates of seed removal (1996 and 1998). Seed removal rates and rates of discovery were greatest at the forest edge and decreased with distance into the old field. The number of seeds surviving to the end of the experiment also varied across the edge gradient, with highest survival at greater distances into the old field in low predation years. Seed removal rate covaried with spatial pattern of Peromyscus leucopus captures within the site. Seed removal and discovery was also higher under the exotic shrub Rosa multiflora, which may have provided cover for foraging seed predators. These indirect effects of edges on plant communities can potentially alter the rate and spatial pattern of tree invasion into disturbed lands and illustrate the importance of understanding plant-animal interactions in the context of habitat fragmentation.     

Large plant size counteracts early seed predation during the extended flowering season of a Silene uniflora (Caryophyllaceae) population

A population of the alvar race of the perennial herb, Silene uniflora (Caryophyllaceae), which grows on thin soils in open alvar habitats on the Baltic island of Oland SE Sweden, was found to have an extended and more or less bimodal flowering phenology Large individuals produced flowers during both periods, while small, and presumably young, individuals only produced flowers in either of the two periods, or in part of both In the early flowering period plants were heavily infested by the seed-predatory larvae of a noctuid moth, Hadena confusa , but in the late flowering period only a small proportion of the fruits was attacked by the seed predators The proportion of flowers developing into fruits was consistently high throughout the season Both the number of seeds per capsule and the mass of seeds decreased over the flowering season However, the germination success of early and late seeds did not differ Thus, although differing in number of seeds, both early and late flowers contributed to individual reproductive success Large individuals started to flower early in the season and despite their high loss of seeds in the early part of the summer they contributed a larger number of seeds to the seed pool than smaller and later-flowering individuals Although selection was acting to favour later flowering during a year with high early seed predation, consistency of date of flower initiation and of relative predation impact on individuals across years was low suggesting that recurrent selection by seed predators is weak Seed predation, although heavy, is therefore judged to be unlikely to cause a significant evolutionary response on flowering phenology in this plant     

Fire and dynamics of granivory on a California grassland forb

This study examines the effects of burning and granivory on the reproductive success of the rare plant Amsinckia grandiflora (Boraginaceae). Fire is often used in California grasslands as a means of exotic species control, but the indirect effects these controls have on the reproductive ecology of native plants are rarely assessed. The interaction of fire with granivory of A. grandiflora seeds was examined in California grasslands over five years (1998–2002). In 1998 and 1999, both burned and unburned plots had bird-exclusion (netted) and no-exclusion (open) treatments. Predation rates were high (51–99%) and final predation rates did not differ among treatments. In 2000, granivory rates in the unburned, open plots were lower than in previous years (14%), and rodent trapping yielded only a single animal. Low granivory rates were observed in 2001 for unburned, open plots (47%). In 2001, burned/open plots experienced significantly more granivory (87%) than either burned/netted plots (37%) or unburned/open plots (47%). In 2002, every seed was taken from burned, open plots. Granivory was highly variable, ranging from 4 to 100% per plot over a 3-week period. Nearly all plots were discovered (>10% predation) by granivores in all trials in all years. When data from all treatments were combined, significant differences in granivory rates occurred among years, indicating stronger inter-year effects than within-year effects due to burning or bird exclusion. Fire affects granivory when overall predation rates are low, but when predation levels are high (as they were in 1998 and 1999), fire may not affect granivory occurring within the same year. Models extending seed survivorship through the dry summer indicate that most seeds are eaten, even when granivory rates are low.     

Seed predation,pathogen infection and life-history traits in Brassica rapa

Herbivory and disease can shape the evolution of plant populations, but their joint effects are rarely investigated. Families of plants of Brassica rapa (Brassicaceae) were grown from seeds collected in two naturalized populations in an experimental garden. We examined leaf infection by the fungus Alternaria, seed predation by a gall midge (Cecidomyiidae) and plant life-history traits. Plants from one population had heavier seeds, were more likely to flower, had less fungal infection, had more seed predation and were more fecund. Fungal infection score and seed predation rate increased with plant size, but large plants still had the greatest number of undamaged fruits. Spatial heterogeneity in the experimental garden was significant; seed predation rate and fecundity varied among blocks. An apparent tradeoff existed between susceptibility to disease and seed predation: plants with the highest fungal infection score had the lowest seed predation rate. Alternaria infection varied between populations, but the disease had no effect on fecundity. Seed predation did reduce fecundity. Damaged fruits had 31.4% fewer intact seeds. However, evidence for additive genetic variation in resistance to seed predation was weak. Therefore, neither disease nor seed predation was likely to be a strong agent of genetically based fecundity selection.     

Phenological variation in fruit characteristics in vertebrate-dispersed plants

Summary We investigated inter-specific variation in fruit characteristics — fruit size, seed number per fruit, seed weight, nutritional content, fruit persistence, and fruit synchronization — in relation to flowering and fruiting phenology in 34 species of fleshy fruited plants. Except for aspects of fruit synchrony and persistence, the results in general were inconsistent with previous suggestions about adaptive variation in phenologically related fruit traits. The main results were as follows: (1) Late flowering, late fruiting, lengthy development time from flower to fruit, and highly persistent fruits constitute a complex of correlated characteristics among the species. (2) Synchronization of fruiting within individuals increased from early ripening fruits to late ripening fruits. Fruiting synchrony was more pronounced in species with a small crop size than in species with a large fruit crop, whereas synchrony was not significantly related to flowering synchronization, nor to life form. (3) Nitrogen and carbohydrate content of fruit pulp did not vary in relation to phenology, whereas lipid content decreased from early to late ripening fruits. (4) No seasonal trends were found for variation in seed size or seed number per fruit. (5) Interactions with flowering phenology and developmental constraints are important in phenological fruiting patterns. Temporal variation in start of fruiting was partly (36%) explained by variation in flowering time. Seed weight variation explained 17% of variation in development time from flower to fruit. (6) Despite constraints from flowering and seed development, some adaptive adjustment in fruiting phenology is likely to be allowed for among the investigated species. Such an adaptive variation in fruiting phenology was suggested by intra-generic comparisons of Prunus and Vaccinium species.     

Density-dependent flowering phenology,outcrossing, and reproduction in Impatiens capensis

Summary We investigated the effects of plant density on cleistogamous (CL) and chasmogamous (CH) flowering phenology and seed production in a natural Impatiens capensis population, by censusing individually marked plants at experimentally reduced and natural densities. CL flowering was earlier at natural density. This plastic density response may have resulted from a stress-related threshold for CL flowering; slower growing plants at natural density flowered earlier. Although apparently triggered by slow early growth, early CL flowering also involved an additional cost for later growth rate. In contrast, CH flowering was unrelated to relative growth rate, but apparently required a size threshold. Experimental density reduction resulted in earlier CH flowering and a dramatic increase in the percentage of plants producing CH flowers. Individual CL and CH flowering duration and flower production were greater at reduced density. These density-dependent effects caused differences between treatments in the shape and location of population flowering phenology curves. Moreover, the percentage of CH seeds produced per individual was much higher at reduced density. At natural density total seed production per plant was lower and more hierarchical than at lower density, suggesting that dominance and suppression shape jewelweed fitness distributions.     

Genetic variation in flowering phenology and avoidance of seed predation in native populations of Ulex europaeus

The genetic variation in flowering phenology may be an important component of a species’ capacity to colonize new environments. In native populations of the invasive species Ulex europaeus, flowering phenology has been shown to be bimodal and related to seed predation. The aim of the present study was to determine if this bimodality has a genetic basis, and to investigate whether the polymorphism in flowering phenology is genetically linked to seed predation, pod production and growth patterns. We set up an experiment raising maternal families in a common garden. Based on mixed analyses of variance and correlations among maternal family means, we found genetic differences between the two main flowering types and confirmed that they reduced seed predation in two different ways: escape in time or predator satiation. We suggest that this polymorphism in strategy may facilitate maintain high genetic diversity for flowering phenology and related life‐history traits in native populations of this species, hence providing high evolutionary potential for these traits in invaded areas.     

VARIATION IN INDIVIDUAL FLOWERING TIME AND REPRODUCTIVE SUCCESS OF AGALINIS STRICTIFOLIA (SCROPHULARIACEAE)

Field studies on two populations of Agalinis strictifolia were conducted over a 3-year period to investigate the relationship between flowering time of individuals and plant size, flowering duration, flower and fruit production, fruit predation, and growth rate. Seasonal patterns of pollinator visitation were compared with those of individual flowering time, flower density, percent fruit production, and mean seeds/fruit. In general, early and middle flowering individuals (as determined by either first flowering date or peak flowering) were larger, flowered longer, and produced more flowers and fruits than late flowering individuals. Early and middle flowering individuals (based on first flowering date) also grew faster than late flowering individuals. Although early and middle flowering individuals produced more fruits, fruit predators did not damage a disproportionate number of fruits compared to late flowering individuals. Patterns of bee visitation showed no association with seasonal patterns of flower density, percent fruit production, mean seed/fruit, or individual flowering time. In populations of A. strictifolia, it would seem that biotic or environmental determinants of growth rate (hence size and reproductive success) may be more important in generating variation in individual flowering time than patterns of pollinator visitation or fruit predation.     

Population dynamics of the perennial Plantago media in semi‐natural grasslands

Abstract. In this study we investigated population dynamics of a perennial forb, Plantago media, in semi‐natural grasslands in southeastern Sweden. Plantago media is a rather common inhabitant of grasslands in this region, but it has been shown to experience dispersal limitations both among managed grassland sites and at potential dispersal routes along road verges. The demography of P. media was studied in 22 permanent plots at two sites over five years. A seed sowing experiment was also performed at each site. The life‐cycle of P. media includes seedlings, juveniles, small vegetative rosettes, large vegetative rosettes and flowering rosettes. The population growth rates (λ‐values) were negative for both populations, with one exception: the first year transition matrix at one site. The projected life span for individuals surviving from seedlings to flowering rosettes varied between 19.6 yr and 227.8 yr. Elasticity analysis showed that remaining in the large rosette stages, both vegetative and flowering, was the most important factor influencing population growth rate. LTRE analysis gave similar results, indicating that spatial and temporal variation have similar effects on the population growth rate. The expected time to extinction in populations with an initial size of between 100 and 1000 individuals varied between 60 and 200 yr. The seed sowing experiment showed that seedling emergence was enhanced by both seed addition and disturbance, suggesting that recruitment is limited by seed and microsite availability. Recruitment was not found to have much impact on population growth rate. However, long‐term population maintenance may depend on occasional small‐scale disturbances that enhance pulses of recruitment. Fragmented landscape is likely to effect plant populations including rare and endangered species as well as more common species which have limited dispersal mechanisms.     

On the density-dependence of seed predation in<Emphasis Type="Italic"> Dipteryx micrantha</Emphasis>, a bat-dispersed rain forest tree

We studied the effect of seed density on seed predation by following the fate of bat-dispersed Dipteryx micrantha (Leguminosae) seeds deposited under bat feeding roosts. The study was conducted in Cocha Cashu biological station, Amazonian Peru, during the fruiting period of Dipteryx. Predation of Dipteryx seeds in the area is mainly by large to medium-sized rodents. Seed deposits beneath bat feeding roosts were monitored for a 13-week period in an 18-ha study area. A total of 210 seed deposits were found, and on average, seed predators encountered 22% of them during any one week. About one-third of the seed deposits escaped predation, and those deposits that had relatively few seeds were more likely to go unnoticed by rodents than were deposits with many seeds. The mean seed destruction rate was 8% per week; deposits with many seeds tended to lose a smaller proportion of their seeds to seed predators than did deposits with few seeds. Regression tests for the weekly data showed that, at the beginning of the observation period, seed predation was not density-dependent. Later, when the total seed crop beneath roosts was high, the number of seeds predated per deposit was positively density-dependent, while the proportion of seeds predated was negatively density-dependent, indicating predator satiation. Seed deposits that had been visited by seed predators once had a higher probability of being revisited the week after, especially if they contained many seeds when first encountered. This indicates that the foraging behavior of rodents may be affected by their remembering the location of seed-rich patches.     

Seed Dispersal of a High Quality Fruit by Specialized Frugivores: High Quality Dispersal?1

Dispersal quality, as estimated by the cumulative effects of dispersal, germination, seed predation, and seedling survival, was examined for Beilschmiedia pendula (Lauraceae) in Monteverde, Costa Rica. I determined the pattern of dispersal by finding seeds deposited by birds, protected the seeds from seed predators with cages to assess germination and seedling survival, and examined seed predation rates with marked seeds. Seed predation, germination, and seedling survival were compared between seeds naturally dispersed by birds and seeds placed at randomly located sites. Approximately 70 percent of seeds dispersed by birds (N= 244) were deposited <10 m from crown edges of fruiting B. pendula trees, although some seeds were dispersed at least 70 m away. Larger seeds were more likely to be dispersed under or close to the parent trees, and larger seeds produced larger seedlings. Seed size was not correlated directly with seedling survival, but larger seedlings at three months were most likely to survive one year. Seed predation by mammals and insects and seedling mortality due to fungal pathogens were concentrated beneath the crowns of parent trees. Seedlings and saplings were more abundant beneath fruiting B. pendula trees, but individuals farther away were taller on average. Thus, dispersal is beneficial for B. pendula, but such benefits appear most pronounced at a small spatial scale; seeds dispersed >30 m from the crown edges actually had a lower probability of survival than those dispersed 10–20 m. Only 10 percent of B. pendula. seeds received high‐quality dispersal in terms of landing in the zone with the highest per seed probability of seedling survival 10–20 m from parental crowns.     

Pre-dispersal seed predation reduces the reproductive compensatory advantage of thrum individuals in Erythroxylum havanense (Erythroxylaceae)

Male-sterility mutations in hermaphroditic species represent the first step in the evolution of gender specialization. Male-sterile individuals commonly compensate the loss of the male function by increasing the number or quality of seeds. Because the magnitude of compensation determines the maintenance of females within populations and the evolution of sexual specialization, plant-animal interactions may affect these evolutionary processes if animals are sensitive to such reproductive asymmetries. Here we explore the effect of seed predation on the reproductive compensation of Erythroxylum havanense, a distylous shrub with morph-biased partial male sterility, during two consecutive years. Seed predation reduced the compensatory advantage of thrums in 1987, but not in 1988. Annual differences in the intensity of seed predation seem to be accounted for differences in the onset and synchrony of flowering. Thus, although seed predators may reduce the compensatory advantage of thrums, their impact is modulated by the environmental cues triggering flowering and insect emergence.