Post-dispersal seed predation and the establishment of vertebrate dispersed plants in Mediterranean scrublands

The post-dispersal fate of seeds and fruit (diaspores) of three vertebrate-dispersed trees, Crataegus monogyna, Prunus mahaleb and Taxus baccata, was studied in the Andalusian highlands, south-eastern Spain. Exclosures were used to quantify separately the impact of vertebrates and invertebrates on seed removal in relation to diaspore density and microhabitat. The three plant species showed marked differences in the percentage of diaspores removed, ranging from only 5% for C. monogyna to 87% for T. baccata. Although chaffinches (Fringilla coelebs) fed on diaspores, rodents (Apodemus sylvaticus) were the main vertebrate removers of seed and fruit. Two species of ant (Cataglyphis velox and Aphaenogaster iberica) were the only invertebrates observed to remove diaspores. However, the impact of ants was strongly seasonal and they only removed P. mahaleb fruit to any significant extent. While removal of seed by rodents was equivalent to predation, ants were responsible for secondary dispersal. However, their role was limited to infrequent, small-scale redistribution of fruit in the vicinity of parent trees. Rodents and ants differed in their use of different microhabitats. Rodents foraged mostly beneath trees and low shrubs and avoided open areas while the reverse was true of ants. Thus, patterns of post-dispersal seed removal will be contigent on the relative abundance and distribution of ants and rodents. Studies which neglect to quantify separately the impacts of these two guilds of seed removers may fail to elucidate the mechanisms underlying patterns of post-dispersal seed removal. The coincidence of both increased seed deposition by the main avian dispersers (Turdus spp.) and increased seed predation with increasing vegetation height suggested that selection pressures other than post-dispersal seed predation shape the spatial pattern of seed dispersal. Rather than providing a means of escaping post-dispersal seed predators, dispersal appears to direct seeds to microhabitats most suitable for seedling survival. Nevertheless, the reliance of most vertebrate-dispersed trees on regeneration by seed and the absence of persistent soil seed banks imply that post-dispersal seed predators may exert a strong influence on the demography of the plants whose seeds they consume. Even where microsites are limited, the coincidence of the most suitable microhabitats for seedling establishment with those where seed predation is highest provide a means by which selective seed predators can influence community composition. Received: 19 August 1996 / Accepted: 25 January 1997     

Rodent seed predation and seedling recruitment in mesic grassland

Seedling recruitment of two grasses (Arrhenatherum elatius and Festuca rubra) and two herbs (Centaurea nigra and Rumex acetosa) was measured in areas with and without rodents to which seeds of each species were sown at three seed densities (1000, 10,000 and 50,000 seeds m⁻²) in two seasons (spring and autumn 1995). Seed removal was measured for 10-day periods and the fate of seedlings was followed for 15 months after sowing. The proportion of seed removed ranged from 6 to 85% and increased with increasing seed density for each species. Rodents had no effect on seedling emergence or survival in the spring sowing. In the autumn sowing, rodents reduced seedling emergence of all four species sown at 1000 and 10,000 seeds m⁻² but had no impact at 50,000 seeds m⁻², presumably because of microsite limitation. We suggest the difference between spring and autumn arose because emergence was seed limited in autumn but microsite limited in spring; microsite availability was higher in autumn because a summer drought killed plants, reduced plant biomass and opened up the sward. Fifteen months after the autumn sowing, fewer A. elatius and C. nigra seedlings survived on plots exposed to rodents. This result reflected not only the reduced seedling emergence but also increased seedling mortality (seedling herbivory) in sites exposed to rodents. In contrast, F. rubra and R.acteosa showed density-dependent seedling survival which compensated for initial differences in seedling emergence, so that no effect of rodents remained after 15 months. The results suggest that rodent seed predation and seedling herbivory exert strong effects on seedling recruitment of A.elatius and C. nigra when recruitment conditions are favourable (conditions that lead to high microsite availability) and may contribute to both species being maintained at low densities in the grassland. The results also demonstrate that highly significant impacts of rodent seed predation at the seedling emergence stage can disappear by the time of plant maturation. Received: 2 March 1998 / Accepted: 28 September 1998     

Interspecific competition and insect herbivory reduce bush lupine (Lupinus arboreus ) seedling survival

Seedlings suffer high mortality in most plant populations, with both competition and herbivory proposed as being important mechanisms causing seedling death. The relative strength of these factors, however, is often unknown. Here I ask how interspecific competition for light and insect herbivory jointly affect seedling survival of bush lupine (Lupinus arboreus), a native shrub common to coastal California. Bush lupine seedlings germinate in grasslands during winter, and throughout spring potentially compete for light with surrounding fast-growing annual grasses. By early summer, after grasses have died, seedlings can be defoliated by a locally abundant caterpillar, the western tussock moth (Orgyia vetusta). I examined the relative importance of competition and insect herbivory on seedling survival in two separate experiments. First, I compared seedling mortality in plots either exposed to or protected from tussock moth larvae. Plants were protected from herbivory by the judicious use of insecticide; control plants were sprayed with water. Tussock moth herbivory resulted in significantly greater (31%) seedling mortality. To determine the effects of interspecific competition for light on seedling survival, I manipulated the density of grass surrounding lupine seedlings. I removed all vegetation surrounding some individuals, and left intact vegetation surrounding others. Reducing competition resulted in a 32% increase in seedling survival from February to June, as well as changes in seedling growth. To determine whether there are interactive effects of competition and herbivory on seedling survival, I enclosed tussock moth larvae on half of all surviving seedlings within each of the two prior competition treatments, comparing growth and survival of defoliated and undefoliated seedlings. Defoliation in June led to an additional 50% mortality for individuals that had grown with competitors through spring, and a 53% additional mortality for seedlings that grew without competitors through spring. Thus, although competition and herbivory both caused substantial seedling mortality, there was no statistical interaction between these factors. Competition-free plants were not less vulnerable to herbivory than plants that previously grew with competitors. Taken together, these experiments indicate that competition and herbivory are both important sources of mortality for bush lupine seedlings. Received: 4 April 1996 / Accepted: 5 November 1996     

Effects of harvester ants on plant species distribution and abundance in a serpentine grassland

Seed harvesting ants can have important effects on the composition and structure of plant communities. We investigated two effects of Messor andrei, the black seed-harvesting ant, on a serpentine grassland plant community in northern California. First, to determine if selective seed predation by ants affects plant community composition, we excluded harvester ants from 1-mediameter circular plots of grassland. Abundances of all species on these plots and on control plots were measured before and after exclosure. Second, to determine if M. andrei nest mounds affect plant community composition, we compared plant species abundances on and off nest mounds. M. andrei deposit large amounts of organic matter on their nest mounds over a foraging season, so mounds may alter the edaphic environment. The exclusion of seed-harvesting activity did not cause changes in the plant community. Nest mounds had a strong effect on plant communities: there were many more grasses and fewer forbs on ant mounds, although at least one forb, Lepidium nitidum, produced twice as many seeds when it grew on nest mounds. We found that nest mounds formed islands of higher-temperature soil in the serpentine grassland. Received: 31 March 1997 / Accepted: 6 May 1997     

Post-dispersal predation on Pinus sylvestris seeds by Fringilla spp: ground substrate affects selection for seed color

Background matching might lower the risk of seeds being eaten by seed predators that search visually. In aviary experiments, we analyzed the selection of diff erent-colored seeds by ground-feeding finches (Fringillacoelebs and F.montifringilla) against four naturally occurring forest soil substrates. The substrates were fresh burn (black), 6-year-old burn (brown), mineral soil (pale yellow) and Pleuroziumschreberi feather moss (green). We used color-sorted seeds of Pinussylvestris, a species with a large natural variation in seed color, ranging from pale yellow to black. Although seeds were scattered on the substrates at a density of only 91 seeds m⁻², birds removed seeds effectively. Both bird species found more pale than dark seeds on the fresh burn substrate. F. montifringilla also recovered more pale than dark seeds on the old burn, and more dark than pale seeds on mineral soil. In moss, the birds found very few seeds compared to the other substrates, and there was no color discrimination. P.sylvestris is frequently regenerating after fire, suggesting that dark seeds would be favored under selection from visually searching predators. Fire-adapted conifers with serotinous cones, e.g., Pinuscontorta ssp. latifolia, which spread their seeds primarily on freshly burnt surfaces, produce uniformly black or dark brown seeds. However, regeneration of the non-serotinous P.sylvestris is often extended for several years after a fire, during which substrate color and structure change. This may have helped to maintain variation in seed color. When regeneration of a plant species occurs on a substrate of uniform color, we believe that selection by visually searching seed predators will result in the evolution of cryptic seed color. Received: 16 August 1996 / Accepted: 30 November 1996     

Seed dispersal in Cakile edentula var. lacustris : decoupling the fitness effects of density and distance from the home site

A factorial design of three densities of siblings at three local distances from seed parents was employed to distinguish effects of density from effects of dispersal distance on lifespan and fruit production of Cakile edentula var. lacustris, a plant with heteromorphic seeds. The segmented fruits produce two seed types: proximal and distal, with distal seeds having greater mass and greater dispersibility. Effects of longer distances (0.5 km and 30 km) on lifespan and fruit production were investigated using plants at low density. The prediction was tested that the greater seed mass of distal seeds increases fitness when seeds are dispersed into sites of unknown quality away from the home site or when seeds are dispersed to low density. High density caused earlier mortality and lower probability of reproduction. Distance from the maternal plant did not influence lifespan or reproduction at distances of 15 m or less, but lifespan was longer 0.5 km from the home site. No interaction was detected between the effects of density and distance on either lifespan or total fitness. Environmental conditions that influence fitness did not vary as a function of dispersal distance in this system, and favorable conditions at the home site did not persist between generations. Therefore, selection on dispersion patterns in natural conditions is likely to be through effects of density rather than dispersal distance. Proximal seeds had greater reproduction than distal seeds at the home site, and distal seeds had greater reproduction at the more distant sites (but not the most distant site), as expected, but these performance differences could not be attributed to differences in mass between the two seed types. Reduced seed mass was favored at the most distant site, but larger seed mass was favored most strongly at low density. Seeds that are dispersed to low density are larger, suggesting that although kin selection may limit the effectiveness of individual selection to increase seed mass under conditions of sibling competition, density-dependent individual selection on seed mass, rather than distant-dependent selection, also contributes to the observed associations among seed type, seed mass and dispersal ability. Received: 21 October 1996 / Accepted: 4 December 1996     

Functional extinction of a desert rodent: implications for seed fate and vegetation dynamics

Population declines of once‐abundant species have often preceded understanding of their roles within ecosystems. Consequently, important drivers of environmental change may remain undiagnosed because we simply do not know how species that are now rare or extinct shaped ecosystems in the past. Australia's desert rodents are thought to have little numerical impact on seed fate and vegetation recruitment when compared with ants or with desert rodents on other continents. However most research on granivory by Australian desert rodents has occurred in areas where rodents were rare or functionally extinct. Here we ask if the paradigm that rodents are relatively un‐important granivores in Australian deserts is an artefact of their historical decline. In the Strzelecki Desert, the endangered rodent, Notomys fuscus is rare where introduced mesopredators are abundant but common where dingoes (an apex predator) suppress mesopredator populations. We used foraging trays to compare rates of seed removal for a common shrub (Dodonaea viscosa angustissima hopbush) between areas where N. fuscus, hopbush shrubs and their seedlings were rare and common and found that seed removal was consistently higher where rodents were common and hopbush rare. By excluding ants and rodents from foraging trays we show that ants removed more seeds than rodents where rodents were rare but rodents removed far more seeds than ants where rodents were common. By manipulating rodents’ access to the soil seed‐bank we show that hopbush seeds persisted in greater numbers where rodents were excluded than where they had access. Our results support the hypothesis that granivory by rodents may once have been a far more important process influencing the fate of seeds and shaping plant communities in arid Australia and suggest that dingo extirpation has cascading effects on shrub seeds. Our study highlights that functional extinction of rodents may be an under‐appreciated driver of vegetation change.     

Seed bank dynamics of the desert cactus Opuntia rastrera in two habitats from the Chihuahuan Desert

In desert environments the main input to the seed bank of many succulents is the seed rain through zoochory while high levels of granivory by rodents, birds and ants are the main cause of subsequent losses. In the patchy environment of arid lands the characteristics of both processes may vary between habitats causing differences in the recruitment of new genets. To test this hypothesis we used populations of the desert cactus Opuntia rastrera which has different recruitment rates in the two adjacent habitats where it grows. In Opuntia-dominated scrublands (nopaleras, density ca. 4,000 plants/ha) 1 seedling out of 7,000,000 seeds establish, whereas in grasslands (density ca. 100 plants/ha) this ratio is 1:20,000. From 1996 until 1998 the seed rain, seed removal by granivores and seed abundance in the soil were monitored in both habitats. Results showed striking differences in the dynamics of the seed bank of both habitats. Seed rain was 8.5 times bigger in nopaleras than in grasslands. In nopaleras most seeds were removed by rodents while the quantities of seeds removed by rodents, birds and ants in grasslands were similar. One year after dispersal (the time necessary to break seed dormancy) only 6% of original nopalera seeds and 12% of grassland seeds remained. After germination trials only 1% (ca. 15,000 seeds/ha) and 2% (ca. 2,500 seeds/ha) respectively were viable. These differences in the effective seed bank (6 times bigger in nopaleras) can not explain the differences in genet recruitment (which is several orders of magnitude bigger in grasslands). Apparently the between habitat difference in nurse plant availability and in rodent density (which inflict a strong hervibory upon seedlings) can explain the differences in genet recruitment. It is speculated that this between habitat difference in genet recruitment suggests that the species evolved in less extreme environments (e.g. grasslands) than desert scrublands which, in turn, are colonised due to the singular ability of O. rastrera for vegetative propagation. This revised version was published online in August 2006 with corrections to the Cover Date.     

Effects of experimental manipulation of inflorescence size on pollination and pre-dispersal seed predation in the hummingbird-pollinated plant Ipomopsis aggregata

Large floral displays should theoretically provide advantages to plants through increased pollinator visitation and resulting fruit and seed set. However empirical tests of the response of pollinators to floral display size have been limited by a lack of direct experimentation, and the results of such studies have been equivocal. In addition, other selective agents such as pre-dispersal seed predators might modulate effects of floral display on pollination. By artificially altering flower number, we examined the direct effects of floral display in the monocarpic herb, Ipomopsis aggregata (Polemoniaceae), on visitation rates by broad-tailed and rufous hummingbird pollinators, as well destruction of fruits by a pre-dispersal seed predator (Hylemya: Anthomyiidae). In addition, we quantified the ultimate effects of flower number on female reproductive success. Plants with larger floral displays were most likely to be visited first in any given foraging bout (P < 0.01). As expected, plants with more flowers received more total flower visits. However, we found no gain in the proportion of flowers visited for many- versus few-flowered plants, or the total number of approaches/hour. In fact, a significantly greater percentage of flowers were visited on few-flowered plants. Plants did not compensate for our reduction in flowers by increasing investment in the number or proportion of flowers that set fruit, the number of seeds/fruit, or seed weight. Pre-dispersal seed predation was greater for many- than for few-flowered plants (P < 0.001), but this did not offset the potential fitness gains of producing large displays. Our data support the hypothesis that large floral displays function primarily in long-distance attraction of pollinators, and enhance maternal success. Received: 21 March 1996 / Accepted: 24 October 1996     

Fruit production in Sorbus aucuparia L. (Rosaceae) and pre-dispersal seed predation by the apple fruit moth (Argyresthia conjugella Zell.)

Variation in fruit production and pre-dispersal seed predation by Argyresthia conjugella was studied in␣four populations of Sorbus aucuparia in northern Sweden.␣The number of infructescences, fruits per infructescence, consumed seeds and developed unattacked seeds per fruit were scored in marked trees from 1984 to 1990. The results showed that the number of fruits produced in each population determined the number of seed predators occurring in the host population, as the yearly number of seed predators was significantly and positively correlated with yearly number of fruits, in all but one population. The seed predators showed a delay in response to variation in number of fruits produced. This lag in response resulted in a large proportion of fruits being attacked and seeds consumed in a bad fruiting year that followed a good fruiting year, and vice versa. The proportion of fruits attacked and seeds consumed was largest in the population showing the greatest between-year variation in fruit production and lowest in the population showing the lowest between-year variation in fruit production. Furthermore, the individuals within the former population were synchronised, while they were not in the latter population. These results contradict one of the possible explanations of mast-seeding, where large synchronised between-year variation is supposed to reduce pre-dispersal seed predation. Instead, differences in attraction of the seed predator to differences in fruit crop size could explain the observed difference in seed predation between the two populations with opposite fruiting patterns. Within each population, irrespective of year, the proportion of fruits attacked and seeds consumed was independent of a tree's fruiting display. Therefore, trees with high fruit production, despite harbouring the largest number of seed predators, produced the largest number of developed seeds in absolute numbers, compared to trees that produced few fruits. Received: 25 February 1996 / Accepted: 30 November 1996     

Early reduction of post-fire recruitment of Pinus nigra by post-dispersal seed predation in different time-since-fire habitats

This study analyses the effects of post-dispersal predation of Pinus nigra seeds on the initial recruitment of this species in areas burned by large wildfires, where P. nigra shows very low regeneration. In three different habitats obtained in a gradient of time since fire in Catalonia (NE Spain), we have evaluated the effects of seed predators (ants, rodents and birds) on post-dispersal seed removal and early seedling establishment of P. nigra by using selective exclosures limiting their access to seeds. Ants were the most efficient seed predator group, followed by rodents and birds. The contribution of each group to overall predation showed large seasonal variations. The first seeds dispersed in winter were mainly predated by rodents, which also registered their highest abundance in this season of the year. In spring, at the end of the natural dissemination period of P. nigra seeds, ants became the major predators, this fact coinciding with their increased abundance. Birds showed the lowest predation values. In the seedling establishment experiment, only in the exclusion treatment of the three predator groups was there initial establishment in all habitats, especially in the recently burned area, where there was seedling establishment in all exclusion treatments. The post-dispersal seed predation by different animal groups and low seedling emergence in the different habitats obtained in this study, together with the low seed availability of P. nigra seeds in burned areas, do not predict a favourable outlook for the natural post-fire recolonization of this species, which might even affect its overall distribution area in the region.     

Widespread seed limitation affects plant density but not population trajectory in the invasive plant Centaurea solstitialis

In some plant populations, the availability of seeds strongly regulates recruitment. However, a scarcity of germination microsites, granivory or density-dependent mortality can reduce the number of plants that germinate or survive to flower. The relative strengths of these controls are unknown for most plant populations and for exotic invaders in particular. We conducted a seed addition experiment with a granivore exclusion treatment in a field setting to explore how these factors interact to regulate populations of the widespread invader Centaurea solstitialis (yellow starthistle) at three study sites across the plant’s range in California. We coupled the experimental approach with observational studies within established C. solstitialis populations to estimate seed rain, recruitment and mortality at natural densities. Seed limitation occurred in both experimental and observational plots in all populations. Although vertebrate granivores were active at each site, they had no effect on C. solstitialis recruitment. Density increased mortality, but the effect was variable and weak relative to its effect on fecundity. The seed limitation that was evident at the seedling stage persisted to flowering. Seed-limited populations such as these ought to be highly sensitive to losses to seed predators, and many biological control agents, including those established for C. solstitialis, are seed predators. However, flowering plant density was decoupled from seed production by a strong compensatory response in the surviving plants; seed production was nearly constant in plots across all seed addition levels. Thus, flowering plant density is reduced by the established biocontrol agents, but seed production compensates to replace the population every generation, and no long-term decline is predicted.     

Post-dispersal seed predation of three grassland species in a plant diversity experiment

Aims Post-dispersal seed predation is an important ecosystem process because it can influence the seed's fate after the initial dispersal from the mother plant and subsequently transform communities. Even at small scales, post-dispersal seed predation can vary greatly depending on seed identity, granivorous taxa or microhabitat structure. However, little is known about the role of plant species richness and functional group richness in post-dispersal seed predation. The overall aim of this study was to test whether increasing plant species richness or plant functional group richness affects the rate and variability of post-dispersal seed predation. We additionally investigated the influence of vegetation structure and seed species identity on the rate and variability of post-dispersal seed predation and whether the influence of different granivorous taxa changed with increasing plant species richness.Methods We conducted seed removal experiments along a long-term experimental plant diversity gradient, comprising plots with monocultures to 60 species mixtures of common grassland species in Jena, Germany, in August 2011. We studied seeds of Onobrychis viciifolia, Pastinaca sativa and Trifolium pratense in exclusion experiments (seed cafeterias), an experimental setup that allowed access either for arthropods or slugs or for all granivorous taxa. Traditionally, seeds removed from seed cafeterias were classified as consumed but we used traceable fluorescent-coloured seeds to obtain more accurate predation rates by subtracting recovered seeds from overall removed seeds. The effect of multiple vegetation variables on mean and variability of seed predation rates was analysed using generalized mixed-effect models and linear regressions, respectively.Important findings Rates of recovered seeds were low but contributed to significant differences between seed predation rates and removal rates of seeds in some treatments. Seed predation rates were not directly correlated with increasing plant species richness or plant functional group richness but were influenced byseed species identity and granivorous taxa. Vegetation variables such as vegetation height and cover were significantly associated with seed predation rates. Depending on the seed species and/or the granivorous taxa, different vegetation variables correlated with seed predation rates. Our results indicate that effects of plant functional group richness and multiple vegetation variables on the magnitude of post-dispersal seed predation varied with seed identity and seed predator taxa. A direct effect of plant species and plant functional group richness could be shown on the variability of post-dispersal seed predation for some seed species and their respective predators. Thus, the changes in magnitude of post-dispersal seed predation with increasing plant species richness could potentially impact the fitness of some plant species and thereby influence plant community structure.     

Wind-dispersed seed deposition patterns and seedling recruitment of Artemisia halodendron in a moving sandy land

* BACKGROUND AND AIMS: Artemisia halodendron is a native sub-shrub that occurs mainly in moving and semi-fixed sandy lands in Inner Mongolia, China. Information on the spatial patterns of wind-dispersed seed deposition and seedling recruitment of A. halodendron inhabiting moving sandy lands is very limited. The aim of this study was to examine wind-dispersed seed deposition patterns and post-dispersal recruitment of A. halodendron seedlings. * METHODS: The spatial patterns of wind-dispersed seed deposition and seedling recruitment of A. halodendron were examined by investigating the numbers of deposited seeds, emerged and surviving seedlings using sampling points at a range of distances from the parent plant in eight compass directions for two consecutive growing seasons. * KEY RESULTS: Wind-dispersed seed deposition showed considerable variation between directions and years. Wind transported A. halodendron seeds only a few meters away from the parent plant in all eight directions. Seedling emergence and establishment also showed between-direction and between-year variability, but the spatial pattern of seedling distribution differed from that of seed deposition. Only a very small fraction (<1 %) of the deposited seeds emerged in the field and survived for long enough to be included in our seedling censuses at the end of the growing season. * CONCLUSIONS: The spatial variation in wind speed and frequency strongly affects the pattern of seed deposition, although the variation in seed deposition does not determine the spatial pattern of seedling recruitment. Seeds of A. halodendron are not dispersed very well by wind. The low probability of recruitment success for A. halodendron seedlings suggests that this species does not rely on seedling recruitment for its persistence and maintenance of population.     

Harvester ants modify seed rain using nest vegetation and granivory

1. In annual plant communities, the seed bank determines each generation's potential emergence and any factor that influences it could alter community structure. Harvester ants are common seed predators that may cause seed sources and seed rain composition to vary. 2. In semi‐arid shrubland of the Negev desert of Israel, we hypothesised that the positive effects of dense vegetation on nests of the harvester ant Messor ebeninus Sant. would outweigh reduction by granivory. 3. To test whether nests were seed sources, we removed vegetation on 10 of 20 sampled nests prior to seeding. We tested the interaction between seed sources and granivory by locating seed traps at three locations: nests, foraging trails, and unvisited areas. Trapped seeds were recorded during spring and summer dispersal. Plant recruitment from the altered seed bank was recorded the following spring. 4. During April, seeds were more abundant near nests than the other locations and at uncut compared with cut nests. Foraging trails had fewer species and higher equitability compared with nests or unvisited areas. May and June to August had highest species richness at nests, but richness of cut and uncut nests were similar. Ordination of plant recruitment showed slightly lower vegetation community variation by diminished species in the cut seed bank. 5. While intact nest vegetation initially had high abundance of locally deposited seeds, it apparently blocked wind‐dispersed seeds later in the summer. Conversely, granivory reduced seed species richness only on trails and decreased dominance structure by preferential removal of some species.     

Effects of dung and seed size on secondary dispersal,seed predation,and seedling establishment of rain forest trees

Seeds dispersed by tropical, arboreal mammals are usually deposited singly and without dung or in clumps of fecal material. After dispersal through defecation by mammals, most seeds are secondarily dispersed by dung beetles or consumed by rodents. These post-dispersal, plant-animal interactions are likely to interact themselves, as seeds buried by dung beetles are less likely to be found by rodents than unburied seeds. In a series of three experiments with seeds of 15 species in central Amazonia (Brazil), we determined (1) how presence and amount of dung associated with seeds influences long-term seed fate and seedling establishment, (2) how deeply dung beetles bury seeds and how burial depth affects seedling establishment, and (3) how seed size affects the interaction between seeds, dung beetles, and rodents. Our overall goal was to understand how post-dispersal plant-animal interactions determine the link between primary seed dispersal and seedling establishment. On average, 43% of seeds surrounded by dung were buried by dung beetles, compared to 0% of seeds not surrounded by dung (n=2,156). Seeds in dung, however, tended to be more prone than bare seeds to predation by rodents. Of seeds in dung, probability of burial was negatively related to seed size and positively related to amount of dung. Burial of seeds decreased the probability of seed predation by rodents three-fold, and increased the probability of seedling establishment two-fold. Mean burial depth was 4 cm (0.5–20 cm) and was not related to seed size, contrary to previous studies. Probability of seedling establishment was negatively correlated with burial depth and not related to seed size at 5 or 10 cm depths. These results illustrate a complex web of interactions among dung beetles, rodents, and dispersed seeds. These interactions affect the probability of seedling establishment and are themselves strongly tied to how seeds are deposited by primary dispersers. More generally, our results emphasize the importance of looking beyond a single type of plant-animal interaction (e.g., seed dispersal or seed predation) to incorporate potential effects of interacting interactions.     

Directed vertebrate-mediated seed dispersal of a fleshy-fruited amaryllid in a heterogenous habitat

Most plants with fleshy fruits have seeds that are ingested by animals, but a less well-understood mode of seed dispersal involves fleshy fruits containing seeds that are discarded by frugivorous animals because they are too large or toxic to be ingested. We studied the seed dispersal biology of Haemanthus deformis, an amaryllid lily species found in a mosaic of bush clumps in a grassland matrix in South Africa. We asked whether seed dispersal is directed in and among bush clumps and whether germination and survival are greater for seeds dispersed to bush clumps than for those dispersed into grassland. Using camera trapping, we found that fruits are consumed mainly by birds and rodents. The pulp was removed from the seeds which were then discarded without ingestion. While many seeds were dispersed close to the parent plant, most (c. 78.5%) were dispersed further than 1 m away from the parent plant. Longer distance dispersal resulted mainly from birds flying off with fruits in their bill or from rodents engaging in scatter-hoarding behavior. Seedling survival was most successful within bush clumps as compared to grasslands and shade was identified as a primary requirement for seedling survival. Seeds from which the fruit pulp had been removed germinated faster than those in intact fruits. Haemanthus deformis deploys a system of directed seed dispersal, whereby both birds and rodents contribute to the dispersal of seeds within patchy bush clumps that are favorable for seedling survival.     

Life on the edge: conspecific attraction and recruitment of populations to disturbed habitats

Variation in the recruitment of benthic marine invertebrates is often attributed to the interaction of the supply of new individuals to a habitat and the availability of space for colonisation when they arrive. Also important in determining variation in recruitment is the response of the larvae to the characteristics of the habitat. Larvae of many benthic marine invertebrates have shown great specificity of requirements in setting their limits of distribution at the time of selection of a habitat. The tubeworm Galeolaria caespitosa shows great variation in recruitment from place to place on rocky intertidal seashores and is a gregarious animal with larvae showing directed responses to conspecific adults on the substratum. I hypothesised that, if variation in recruitment of G. caespitosa were independent of conditions on the substratum, the magnitude of recruitment in patches of the same shape but different sizes cleared within continuous mats of conspecific adults would be directly related to the area available for colonisation in the patch. Alternatively, if variation in recruitment were due to the response of larvae to conspecific adults on the substratum, the magnitude of recruitment would be a function of the perimeter of the patch, which, given patches of the same shape, is a measure of the influence of conspecific adults in that patch. To distinguish between these alternatives, small (area = 25 cm²; perimeter = 20 cm) and large (area = 225 cm²; perimeter = 60 cm) square patches were cleared within continuous mats of conspecific adults at four sites and recruitment of G. caespitosa was monitored over two seasons of recruitment. The density of recruits per unit area was, on average, almost three times greater in small than in large patches, indicating that recruitment of G. caespitosa is not directly related to the area of the patch. In contrast, the density of recruits per unit perimeter was not significantly different between small and large patches, indicating that recruitment of G. caespitosa is related to the proximity of conspecific adults in the patch. Therefore, at a given site, the perimeter of patches within mats of G. caespitosa is a better predictor of the relative magnitude of recruitment among patches than that provided by their areas. These results are contrary to many models of invertebrate recruitment that assume close linkage between available space on the substratum and settlement. Moreover, they highlight the importance of behavioural responses of animals at the time of selection of habitat in accounting for variation in recruitment. For populations of organisms that display gregarious behaviour at settlement, or conspecific attraction, this direct relationship between the perimeter of patches and recruitment could be used as a tool in restoring populations to disturbed habitats. The added benefit of such facilitative interactions in restoring populations is that they provide increasing returns to the population for a given supply of potential colonists to a habitat. Received: 1 November 1996 / Accepted: 20 January 1997     

Population ecology of wild sunflowers: effects of seed density and post-dispersal vertebrate seed predators

Assessing the effects of seed density on the population dynamics of wild plant species with crop relatives will be vital in determining the potential effects of introducing traits into wild populations as a result of crop-to-wild gene flow. We examined experimental sunflower (Helianthus annuus) patches in eastern Kansas to determine the effects of seed density and predation on seedling recruitment and seed production in the next generation. High seed density treatment plots had significantly more seedlings and adult plants than did low seed density treatment plots. Overwinter vertebrate seed predator exclusion treatments resulted in increases in plant density compared to plots in which vertebrates were not excluded. Control patches (no seeds added) contained virtually no plants. Head production and estimated total seed production for a patch were not statistically different among treatments (excluding control plots). Although initial seed density and vertebrate post-dispersal seed predation do appear to have effects on seedling recruitment, neither appear to be limiting seed production of competing adult plants. Therefore, variation in seed densities (over the range examined) may have limited effects on local population dynamics. It is important to note that the choice of seed densities may affect the results obtained: the seed densities used in this study may, in retrospect, be higher than in the small roadside populations typical in eastern Kansas, yet other natural sites have much larger densities. Further, the effects of increased seed density at a local site may have other important effects such as altering metapopulation dynamics through increased long-distance dispersal or increased local seed bank size.