Oceanic islands, due to their evolutionary history and isolation, play a dual role of having high endemicity and being vulnerable to extinctions, with most known extinctions occurring on islands. Plant–animal interactions are particularly important on islands, as island systems generally have low redundancy and are more vulnerable to disruption either via extinction or by invasive species. Here, we examined the fruit removal and seed predation of a keystone palm, Caryota mitis, on Narcondam, a remote oceanic island. The island endemic Narcondam Hornbill (Rhyticeros narcondami) was the sole seed disperser of the palm (90 hours; N = 15 trees), with mean (± SE) visitation rate being 0.23 (± 0.06) individuals per hour and fruit removal rates of 3.5 (± 1.5; range: 0–16) fruits per visit, indicating a lack of redundancy in seed dispersal of the palm on this island. Whereas the invasive rodent, Rattus cf. tiomanicus, was the sole predator of palm seeds (N = 15 individual fruiting palms, 416 trap nights). Overall, 17.1% of the seeds placed (N = 375 seeds) were removed. Seeds placed under and away from the canopy, and at different densities (2 plots with 10 seeds each; 1 plot with 5 seeds, respectively), showed similar removal rates. This indicates density-independent seed predation and the lack of safe regeneration sites for Caryota mitis, with potential deleterious effects on subsequent stages of the ‘seed dispersal cycle’. Here, from a data-deficient site, we provide baseline information on the plant–frugivore interaction of a keystone palm and the potential impacts of an invasive rodent.

One of the biggest issues in plant ecology is determining the interaction outcome between seeds and scatter-hoarding rodents because the latter has a dual role as dispersers and predators of seeds. Density-dependence contexts involving resource abundance largely influence the outcome of this interaction. Here, we investigated how the variation in the density of a large-seeded tropical tree (Joannesia princeps Vell) affects its probability of seed removal, consumption, dispersal, and burial by a neotropical rodent (Dasyprocta azarae Lichtenstein). We tested whether the elevated resource availability in high tree density areas would cause scatter hoarder’s satiation by decreasing seed removal and consumption (predator satiation hypothesis) or increasing seed dispersal and burial (predator dispersal hypothesis). We tracked the fate of 461 seeds in 14 plots with distinct J. princeps abundances inside a large Atlantic Forest fragment. We used spool-and-line tracking and camera trappings to determine seed fate and identify interacting animals. Agouti was the only species removing J. princeps seeds. Tree density benefitted J. princeps by increasing seed dispersal through buried seed but not affecting seed removal and consumption. This result shows how density-dependent contexts, such as tree density, may alter seed fate in seed–rodent interactions supporting future studies aiming to reestablishing seed dispersal functions in Atlantic Forest fragments.

Post-fire flowering (PFF) species resprout, flower and disperse seeds within weeks after fire. This may be an important strategy to recruit new individuals and colonize the gaps opened by fire. The seeds released in the post-fire environment may interact with byproducts derived from plant burning, and the resulting ash may have compounds that can promote the germination of various seeds, particularly those with permeable coats. In the Cerrado ecoregion, PFF is a strategy commonly observed in the species of the ground layer, but their seeds are rarely investigated. So, we examined the quality and the germination of the seeds of 13 species that disperse seeds within 3 months after fire. We estimated the amount of empty, filled and predated seeds for each species, and tested the germination with or without ash. There was a clear separation in seed quality as dicots produced 35–75% filled seeds but grasses <15%. Pre-dispersal predation was only observed for dicots (<10%). Ash stimulated the germination of two out of the three dormant species but inhibited the germination of three non-dormant species. Overall, the seeds produced in response to fire are an important source of genetic variability in an ecosystem that has resprouting as the main persistence strategy. As most species have non-dormant seeds, ash may only be important to stimulate the germination of few PFF species. Even so, ash can be completely washed away by rains before seeds are dispersed and may not have an effect under field conditions.

One of the many ecosystem services provided by dung beetles is that of secondary seed dispersal. This paper experimentally evaluates the effectiveness of this service using the chimpanzee–dung beetle seed dispersal system in Taï National Park, Côte d’Ivoire. The study focussed on the germination rate and success of four species of seeds contained in the faeces of Pan troglodytes verus: Dacryodes klaineana (Pierre) H.J. Lam, Diospyros mannii Hiern, Pycnanthus angolensis (Welw.) Warb., and Uapaca guineensis Muell. Arg. For each species, 600 seeds, half from chimpanzee faeces and half from mother trees, were sown in nurseries at depths of 0, 2, 5, 10, and 15 cm (i.e. 120 seeds per depth; 60 from chimp and 60 from mother trees). After germination, only the seeds of Uapaca guineensis and Diospyros mannii sown at 2 and 5 cm had a >14% rate (between 14.2 and 30.8%) of germination and seedling establishment, regardless of seed origin. An increase in the depth of seed burial appears to negatively affect the probability of seedling emergence and establishment. This study shows that dung beetles have positive impacts on seed fate. However, for these plant species, the chimpanzee’s role is limited to that of primary seed dispersal.

The loss of large frugivores leads to seed dispersal loss and regeneration failure of numerous large-seeded trees near mother trees. Although Janzen–Connell effects are considered as the primary underlying cause, other factors remain understudied. Here, we used a field experiment to test the impact of flesh persistence on the recruitment of two large-seeded Sapotaceae species that lost their dispersers. In the rainforest of Mare Longue (Réunion), we sowed 3840 seeds in a four-factor design: seed treatment (seed cleaning; flesh persistence), canopy closure (understory; gap), year of sowing (01/2018; 11/2019) and species (Labourdonnaisia calophylloides, Mimusops balata). We also used camera traps to evaluate the impact of extant vertebrates. Seed treatment was by far the most influential factor: flesh persistence led to seedling recruitment divided by 3,2 on average, mainly due to failure of germination or seedling emergence. There were also significant variations in recruitment between species, years and canopy closure levels, notably due to the behaviour of the invasive fauna, especially giant snails that could unexpectedly restore recruitment by feeding on fruit flesh. Together, our results demonstrate strongly depleted recruitment due to flesh persistence and the importance of field experiments to understand the processes at work in complex ecosystems with novel plant–animal interactions.

Myrmecochory, the dispersal of seeds with lipid-rich appendages by ants, is a significant ant–plant interaction. Less well understood is the potential for ant dispersal of non-myrmecochorous seeds. Here we investigate ant–diaspore interactions in a West African montane habitat. We combine observation with depot experiments to determine ant species that move diaspores and distance moved across a forest-edge-grassland gradient. We recorded seed cleaning by ants using a bird/mammal dispersed Paullinia pinnata to determine whether seed cleaning improved plant fitness. We found that two out of a total of 17 ant species (Pheidole sp. 1 and Myrmicaria opaciventris) interacted with 10 species of non-myrmecochorous diaspores across nine plant families. Diaspores were from large canopy trees, understorey trees and vines. Both ant species interacted with small (≤0.24 g) and large (≥0.24 g) diaspores. Ants individually moved small diaspores up to 1.2 m and worked together to clean larger ones. Our experiments with P. pinnata showed that ants removed the pulp of 70% of fruit over 5 days. Cleaned seeds germinated significantly faster and produced seedlings with significantly longer shoot length and higher fresh weight than seedlings from intact seeds. Together our results suggest that ant dispersal may be less significant than seed cleaning in Afromontane forests. However, given the decline in vertebrate frugivores across Africa, a small dispersal advantage may become increasingly important to plant fitness.

The use of a corn–earthworm coculture (CE) system is an eco-agricultural technology that has been gradually extended due to its high economic output and diverse ecological benefits for urban agriculture in China. However, the effect of CE on weed occurrence has received little attention. A 5-yr successive experiment (2015 to 2019) was conducted to compare weed occurrence in CE and a corn (Zea mays L.) monoculture (CM). The results show that CE significantly decreased weed diversity, the dominance index, total weed density, and biomass, but increased the weed evenness index. The 5-yr mean number of weed species per plot was 8.4 in CE and 10.7 in CM. Compared with those in CM, the 5-yr mean density and biomass of total weeds in CE decreased by 59.2% and 66.6%, respectively. The effect of CE on weed occurrence was species specific. The mean density of large crabgrass [Digitaria sanguinalis (L.) Scop.], green foxtail [Setaria viridis (L.) P. Beauv.], goosegrass [Eleusine indica (L.) Gaertn.], and common purslane (Portulaca oleracea L.) in CE decreased by 94.5%, 78.1%, 75.0%, and 45.8%, whereas the mean biomass decreased by 96.2%, 80.8%, 76.9%, and 41.4%, respectively. Our study suggests that the use of CE could suppress weed occurrence and reduce herbicide inputs in agriculture.

Seed predation has been suggested to play a crucial role in plant population dynamics and could act as a strong selective force in the evolution of plant traits such as plant phenology, size and number of flowers. The loss of seeds could be particularly threatening for population maintenance of rare plant species. We investigated the influence of seed predators on plant fitness of the rare monocarpic thistle Cirsium decussatum. We tested the following hypotheses: (1) pre-dispersal seed predators reduce the number of dispersed propagules; (2) seed predators select larger inflorescences as oviposition sites; (3) the size of the capitulum is correlated with seed productivity; and (4) seed predators compete for the oviposition sites. We identified ten insect taxa from different taxonomic groups in the capitula of C. decussatum. Terellia longicauda larvae (Diptera: Tephritidae) were the most common. The presence of the flies’ larvae was positively correlated with the predation level in the capitula. Females of T. longicauda selected larger inflorescences for laying eggs that may increase offspring fitness as larger inflorescences promise more food resources. We could not prove the presence of competition between insect species inhabiting the capitula. Our data suggest, however, that T. longicauda, a specialized seminophagous dipteran, is one of the factors that negatively influence the fitness of threatened C. decussatum, lowering significantly the seed pool. Terellia longicauda potentially acts as a strong selective force in the evolution of C. decussatum inflorescence size.

Studies were conducted from 1989 to 1993 in continuous no-tillage and moldboard plow corn fields to describe rates of velvetleaf seed predation with time and with seed density, and to identify principal seed predators. Rates of seed loss from the soil surface averaged 1 to 57% day−1 and were equivalent in the two tillage systems. Predator populations were the same in no-tillage and moldboard plow fields. The predation rate was generally low in winter months, increased in mid-summer, and declined in late summer. In 2 of the 4 yr, predation increased in October and November. The predation rate was described by an exponential decay function of seed density, with high rates of seed loss at low densities and leveling off to a nearly constant level at densities above 600 seeds m−2. Predation was highest where seed access was not restricted, and exclosures of 6.5 and 1.6 cm2 reduced predation up to 15 and 52%, respectively. Mice were important predators in the field. In laboratory feeding studies, the carabid beetle Amara cupreolata, the slugs Arion subfuscus and Deroceras reticulatum, and cutworms (Agrotis ipsilon) consumed imbibed velvetleaf seeds. Amara cupreolata and A. subfuscus were the only predators to damage unimbibed velvetleaf seeds.

Consumption of weed seeds and waste grains by seed predators is an important ecosystem service that helps to regulate weed and volunteer crop populations in many agricultural systems. The prairie deer mouse is found in a variety of sparsely vegetated habitats throughout the central United States and is the dominant vertebrate seed predator in row-crop fields (corn and soybean) in this region. Evaluating the preferences of prairie deer mice for common agricultural weed seeds and waste grain is important to understand the potential ability of native mice to regulate volunteer crops and weed populations. We evaluated winter seed preference of deer mice using cafeteria-style feeding trials presented within row-crop fields in central Indiana and used compositional analysis to compare proportional consumption of seeds from five common agricultural weeds (common ragweed, common cocklebur, common lambsquarters, velvetleaf, and giant foxtail) and two grains (corn and soybean) during overnight feeding trials. Prairie deer mice significantly preferred corn to all other available seed types. Ragweed and soybean were also readily consumed and were preferred over seeds other than corn. Giant foxtail was intermediately preferred. Our results show that prairie deer mice have clear preferences for certain seeds commonly available in row-crop fields; mice likely contribute to reduction of waste grain and some weed seed populations.

Seed-consuming rodents are increasingly recognized for their roles in the maintenance of tree species diversity, by acting on crucial post seed-dispersal processes. Yet, studies examining the extent to which rodents may act as secondary seed dispersers and/or predators in Madagascar's forests are limited. Thread-marking seed experiments were run to address this, using the seeds of two native large-seeded, frugivore-dispersed tree species (Abrahamia thouvenotii and Cryptocarya crassifolia), in disturbed and less-disturbed habitats in the rain forests of Ranomafana National Park, during the dry season. Data show that post-dispersal handling of seeds by rodents (predation and removal) was significantly lower in disturbed than in less-disturbed habitats (1.31–3.78 times lower; n = 2200). Also, seeds were more likely to be predated or left on the forest ground after removal than being larder- or scatter-hoarded: 27–78% of the removed seeds were found on the ground, <12% found in burrows and there was no evidence of scatter-hoarding (n = 132). Based on pictures from camera traps, the native rodent species, Nesomys rufus, was potentially responsible for seed predation and/or removal in the less-disturbed habitat; but no indication of the rodent species active in the disturbed habitat was obtained. The lack of scatter-hoarding suggests a limited role of rodents in secondary seed dispersal in this system. These findings form a preliminary account of the potential roles of rodents in post seed-dispersal processes in Madagascar's forests, but this warrants further study.

Postdispersal processes play an important role in the regulation of weed population dynamics. Experiments were conducted at two locations in Arkansas to understand postdispersal loss of five arable weed species important to this region—barnyardgrass, johnsongrass, pitted morningglory, Palmer amaranth, and red rice—between seed dispersal in autumn and the production of fresh seeds the subsequent autumn. Total seed loss through predation, decay, germination (fatal or successful), and loss in viability was estimated, and the influences of residue level and seed burial depth (near ground vs. 5 cm deep) were also examined. On average, the active (i.e., viable) seedbank proportion in spring (5 mo after dispersal) ranged from 8 to 11% (barnyardgrass), 10 to 11% (johnsongrass), 20 to 23% (pitted morningglory), 4 to 6% (Palmer amaranth), and 5 to 10% (red rice) across the two locations. At 1 yr after dispersal, 0.7 to 1.5% of barnyardgrass, 7 to 8% of johnsongrass, 5 to 9% of pitted morningglory, about 1.5% of Palmer amaranth, and 0.2 to 0.7% of red rice were part of the active seedbank for the two locations. There was no evidence to suggest that establishing a vegetation cover (such as a rye cover crop) after harvest of the main crop could accelerate seed predation. Burial depth did not influence seed decay, but most (45 [pitted morningglory] to 99% [Palmer amaranth]) of the seeds retrieved from the predator feeding stations were found buried in the soil substrate, and thus, not available for most predator species. This suggests that practices that allow weed seeds to lie on the soil surface (such as no-till planting in autumn) are highly valuable in encouraging seed predation. The high levels of seed loss observed in this study indicate that seedbank management should be a vital component of integrated weed management strategies.

To improve understanding of relationships between crop diversity, weed management practices, and weed population dynamics, we used data from a field experiment and matrix models to examine how contrasting crop rotations affect velvetleaf. We compared a 2-yr rotation system (corn–soybean) managed with conventional rates of herbicides with a 4-yr rotation (corn–soybean–triticale + alfalfa–alfalfa) that received 82% less herbicide. In November 2002, a pulse of velvetleaf seeds (500 seeds m−2) was added to 7- by 7-m areas within replicate plots of each crop phase–rotation system combination. Velvetleaf seed, seedling, and reproductive adult population densities, seed production, and seed losses to predators were measured during the next year. Velvetleaf seed production was greater in the 4-yr rotation than in the 2-yr rotation (460 vs. 16 seeds m−2). Averaged over 12 sampling periods from late May to mid-November 2003, loss of velvetleaf seeds to predators also was greater in the 4-yr rotation than in the 2-yr rotation (32 vs. 17% per 2 d). Modeling analyses indicated that velvetleaf density in the 4-yr rotation should decline if cumulative losses of seeds produced in the soybean phase exceeded 40%. Achieving such a level of predation appears possible, given the observed rates of velvetleaf seed predation. In addition, no tillage occurs in the 4-yr rotation for 26 mo after soybean harvest, thus favoring seed exposure on the soil surface to predators. Models that included estimates of seed predation indicated that to prevent increases in velvetleaf density, weed control efficacy in soybean must be ≥ 93% in the 2-yr rotation, but could drop to 86% in the 4-yr rotation. These results support the hypothesis that diverse rotations that exploit multiple stress and mortality factors, including weed seed predation, can contribute to effective weed suppression with less reliance on herbicides.

Because herbicide and cultivation efficacy is generally density independent, seedling density following these weed control practices will be proportional to the density of germinable seeds in the seedbank. Most farmers would therefore benefit from management practices that reduce seed inputs, increase seed losses, and reduce the probability that remaining seeds establish. Germination, predation, and decay are the primary sources of loss to the seedbank that may respond to management. Farmers have long prepared stale seedbeds in which shallow soil disturbance encourages germination losses. Postdispersal seed predation by vertebrate and invertebrate granivores may cause high rates of seed mortality in a wide range of cropping systems, but seed dispersal asynchronous with predator activity, and seed burial, may limit the overall effect on the seedbank. Although seeds would seem to be an ideal carbon source for soil microorganisms, limited evidence from a study of wild oat suggests that decay may be less responsive to management than germination, and likely predation. A final management objective, supporting a program that aims to reduce seedbank inputs and increase losses, is to reduce the size of the effective seedbank through manipulation of residues and disturbance to reduce the probability of establishment. Incorporation of green manures generally reduces weed establishment, whereas larger-seeded or transplanted crops may better tolerate the residue-mediated changes in the chemical, biological, and physical properties of the soil surface environment. Evidence from no-till systems further support the hypothesis that changes in soil surface conditions may regulate the abundance of “safe sites” for weed establishment, thereby modulating the size of the effective seedbank.

Temperate and boreal forests in Canada and the northeastern United States have been invaded by several exotic species, including European earthworms (family Lumbricidae) and garlic mustard. Earthworms and garlic mustard co-occur and are both known to adversely impact some native plant species. However, relatively little is known about potential interactions between these two invaders. In a series of growth chamber experiments, we determined the palatability of garlic mustard and six native herbaceous forest species (shooting star, columbine, wild geranium, sweet cicely, butterfly milkweed, and yellow jewelweed) to the common nightcrawler. We also assessed the ability of the common nightcrawler to bury and digest garlic mustard and wild geranium. When offered seeds from garlic mustard and a native plant species, the earthworms ingested more garlic mustard seeds than seeds from four of the six native species. In a mesocosm experiment, the common nightcrawlers apparently digested 72 and 27% of garlic mustard and wild geranium seeds, respectively, that were placed on the soil surface. No seeds were observed on the soil surface at the end of the experiment but the majority of recovered seeds for both species were found within the top 10 cm (3.94 in). More seeds were recovered in 0- to 10-cm and 31- to 40-cm sections for wild geranium than for garlic mustard. No difference in seed recovery was detected at the other depths. Garlic mustard seed is readily consumed by common nightcrawlers and appears to be preferred over some native plant species suggesting that common nightcrawlers may reduce the size of the garlic mustard seed bank.

Yellow starthistle is an invasive, annual, spiny forb that, for the past 30 yr has been steadily advancing up the Salmon River Canyon in west central Idaho. In 1994, a decision was made to attempt to manage yellow starthistle by establishing a complex of biological control agents in a containment zone where the weed was most dense. Between 1995 and 1997, six species of seedhead-attacking insects were introduced and successfully established. By 1999, the insects had spread through the entire containment zone. Following this dispersal, a rapid buildup of insect populations occurred, and, since 2003, seed destruction has fluctuated around 90%. Vegetation monitoring plots, however, have shown no consistent decline in the overall population of yellow starthistle, indicating that the amount of seed produced is still sufficient to allow full replacement. However, county weed control personnel, who are responsible for surveying and destroying outlying populations of yellow starthistle beyond the containment zone, report that, during this period, the number of new, isolated pockets of yellow starthistle they are finding has dropped dramatically. This case study discusses how the biological control program partially met the objective of managing yellow starthistle by reducing the rate of advance of this weed in the Salmon River Canyon.

Cropping system characteristics affect weed management by altering key demographic rates of weeds, including recruitment, seedling survival, fecundity, and seed survival. To facilitate the design and improvement of cropping systems that limit weed population growth, analytical methods are needed to identify weed management “choke points” (weed life stages that vary in response to management and whose variation strongly affects weed population growth rate). The objectives of this study were to (1) determine whether wheat–red clover green manure can limit giant foxtail population growth rate (λ) in a wheat–corn–soybean crop sequence and (2) identify choke points in the giant foxtail life cycle with respect to the green manure treatment. Demographic data were used to construct a periodic matrix model of giant foxtail population growth in a wheat–corn–soybean crop sequence, with either a wheat sole crop (W) or a wheat–red clover intercrop (R) in the wheat phase. Identification of choke points was accomplished by adapting the life-table response experiment (LTRE) design for retrospective perturbation analysis of the periodic matrix model. The difference in λ (Δλ) between the two treatments was decomposed into contributions from each parameter in each rotation phase of the periodic model. Each LTRE contribution was equal to the product of the sensitivity of λ to changes in a given parameter by the treatment difference in that parameter. Those parameters making large contributions to Δλ represented weed management choke points. Giant foxtail population growth rate in the simulation was more than twice as great in the W treatment (λ = 2.54) than in the R treatment (λ = 1.16). Retrospective perturbation analysis indicated that the proportion of seeds surviving predation in the wheat phase made the largest LTRE contribution (0.55) to Δλ, followed by seedling recruitment in the soybean (0.41) and corn (0.20) phases. By identifying weed management choke points within a given system, retrospective perturbation analysis can target research and management efforts for greater reductions in weed population growth.

Contrasting cover-cropping systems were compared to determine whether fundamental differences in cover-cropping strategies affect weed seed predators and resulting seed predation. We conducted typical “feeding” trials in which 25 seeds of each of six weed species, including velvetleaf, wild mustard, yellow foxtail, common lambsquarters, redroot pigweed, and hairy galinsoga, were placed in the field. Exclosures showed that the majority of seed predation could be attributed to invertebrates: 43% out of a total of 56% seed predation for 11 d in 2002 and 40% out of a total of 58% seed predation for 4 d in 2003. The predominant invertebrate seed predator across all entry points of four cropping systems was a ground-dwelling carabid beetle, Harpalus rufipes, which was more abundant in vegetated treatments, particularly red clover, compared with treatments recently tilled and planted to a fall cover crop. In the absence of vertebrates, H. rufipes activity–density was positively correlated with mean seed predation in 2002 (Spearman ρ = 0.489; P < 0.001) but not in 2003 (Spearman ρ = 0.090; P = 0.504), possibly because of a delay between pitfall trapping and predation assay. The activity–density of invertebrate seed predators measured in these systems and the high level of predation imposed on weed seeds at the soil surface indicate that cover-cropping strategies should consider late-season weed management, which maintains seeds on the soil surface and provides desirable habitat for invertebrate predators.

Postdispersal seed predation by ground beetles may be an important form of biological weed control. Field experiments conducted in 2002 and 2003 determined invertebrate seed predators' ability to detect and remove seeds from different experimental substrates. Predation of wild mustard and common lambsquarters was greater when seeds were presented on sand compared with the two types of synthetic finishing pads; however, predation of velvetleaf, redroot pigweed, and hairy galinsoga was unaffected by substrate. Predation rates were not consistent across all experimental substrates. Estimates of invertebrate predation of common lambsquarters, yellow foxtail, and velvetleaf were greater for seeds offered on sand or synthetic pads than for seeds offered on soil or double-sided tape covered with soil. Although each substrate would be useful to estimate treatment effects on relative predation rates, the weed species by substrate interaction should be considered when comparing predation rates across experiments or when the absolute rate predation is of critical importance.

A field study was conducted to determine the effects of giant ragweed emergence time and population density on corn grain yield, giant ragweed seed production, and giant ragweed predispersal seed losses. When weeds and crop emerged concurrently, hyperbolic regression of percent corn yield loss on giant ragweed population densities of 1.7, 6.9, and 13.8 weeds per 10 m2 gave a predicted loss rate of 13.6% for the first weed per 10 m2 in the linear response range at low densities and a maximum yield loss of 90% at high weed densities. Crop yield loss response to weed density was linear when giant ragweed emerged 4 wk after corn, and the regression coefficient indicated a yield loss rate of 1% per unit increase in weed density. A larger proportion of the variation in corn yield loss was explained by weed density (r 2 = 0.99) than by weed biomass (r 2 = 0.81). There was a positive linear relationship between giant ragweed seed production and weed density at each weed emergence time. When giant ragweed emerged with corn, regression equations for 1997 and 1998 gave a predicted seed rain of 146 and 238 seeds m−2 per unit increase in weed density, respectively. In both years when giant ragweed emerged 4 wk after corn, predicted seed rain was 16 seeds m−2 per unit increase in weed density. Viability of total giant ragweed seed was 56 and 38% in 1997 and 1998, respectively, and was not affected by weed emergence time or weed density. Feeding by insect larvae accounted for 13 to 19% of giant ragweed seed viability losses. Granivorous insects infesting giant ragweed seed were identified as a fruit fly (Diptera: Tephritidae), two weevils (Coleoptera: Curculionidae), and a moth (Lepidoptera: Gelechiidae).