EFFECT OF SEED DIMORPHISM ON THE DENSITY-DEPENDENT DYNAMICS OF EXPERIMENTAL POPULATIONS OF ATRIPLEX TRIANGULARIS (CHENOPODIACEAE)

The importance of seed size and density in determining individual plant performance and plant population dynamics in experimental populations of the halophyte Atriplex triangularis was studied. Two distinct seed morphs—large, light seeds and small, dark seeds—are produced by individual A. triangularis plants. Experimental populations consisting of seed size monocultures (large or small seeds) and seed size mixtures were established at three different densities, and the time of germination, plant size, plant survivorship, and plant fecundity were monitored. Marked variation in time of germination was observed among treatments and between seed sizes, but germination within any given treatment occurred over a five- to ten-day period. Large seeds produced larger plants than small seeds did, and this dichotomy was maintained over the course of the entire experiment. Germination date and seed size interacted such that larger plants grew from seeds which germinated earlier than those which germinated later, regardless of seed size. Germination date had a more pronounced effect than seed size did on plant mortality in high density populations. At high density, large seed monocultures experienced greater mortality than small seed monocultures did, but in seed size mixtures, the mortality was evenly distributed between plants from the two seed sizes. Regardless of density conditions and parentage, large and small seeds were produced in equal proportion by the plants. Total seed production, however, was dramatically affected by plant density, and to a lesser degree by germination date. Although seed size effects alone did not appear to affect directly final plant biomass and fecundity, effects of seed size early in ontogeny may have contributed to differences in fecundity.     

Germination variability and the effect of various pre-treatment on germination in the perennial spurge Euphorbia nicaeensis All.

Patterns of seed germination of the perennial spurge Euphorbia nicaeensis were studied in three populations in south-western Spain. We investigated the variation in seed viability and germination among individuals, and among populations over 2 consecutive years. We also studied if diverse factors such as temperature, acid scarification, darkness or caruncle loss affect the germination of the seeds.

Interindividual and interpopulation variability in seed viability was found. E. nicaeensis seeds are nondormant, so differences in interpopulation viability translate into differences in final germination rates. The germination percentage of the seeds from each population was similar in the 2 years studied. The effects of diverse factors were homogeneous in the two populations studied. Darkness has no effect on seed germination, and ecarunculate seeds germinated in the same proportions as carunculate seeds; this could allow the seeds to germinate in the chambers of deserted anthills. Acid scarification significantly reduced the germination percentage in only one of the populations, but over 50% of the seeds germinated, which could allow herbivores to act as occasional dispersing agents. The seeds that were preheated at 100 °C for 1 and 5 min germinated in the same proportions as the control group. The seeds that were preheated at 120 °C for 5 min displayed a significant decrease in germination, but the percentage was over 40% for both populations, indicating that the seeds could still germinate after the passage of a fire.     

Population dynamics of the shrub Acacia suaveolens (Sm.) Willd.: Seed production and predispersal seed predation

Seed production and predispersal seed predation in the shrub Acacia suaveolens were examined over 3 consecutive years in eight populations in south-eastern Australia. Seed-crop sizes varied both between and within populations of different ages. Seed production was maximal in the first one to four flowering seasons after establishment, and then declined with plant age. The size of the annual seed-crop was also influenced by rainfall for that year. Predispersal seed predation varies between populations over fruiting seasons with the initial large seed-crops resulting in predator satiation. Within one fruiting season, no significant variation in the extent of predispersal seed predation was found in any of three populations studied. There were two major forms of predispersal seed loss: toss of whole fruits to Melanterius corosus (Coleoptera: Curculionidae) and external insect seed grazers, and loss of individual seeds within fruits to M. corosus. Exclusion experiments showed that seeds lost to these predispersal seed predators would otherwise have been matured by the parent plant. The effects of predispersal seed predation cannot be directly related to seedling recruitment. Indirectly, such predation may influence the dispersion of seeds in the soil profile and hence, subsequent recruitment.     

Adaptation to habitat in Aquilegia species endemic to Sardinia (Italy): Seed dispersal,germination and persistence in the soil

Abstract

The autecology of the Sardinian endemics Aquilegia barbaricina Arrigoni et Nardi and A. nugorensis Arrigoni et Nardi were investigated. Peaks of anthesis and seed dispersal were recorded for five populations occurring in two distinct habitats, one riparian and one rupicolous. Germination tests were carried out on seed lots belonging to each population by sowing seeds at 10, 15, 20, 25 and 25/15°C. In addition, seeds were incubated for 2 months at either 25°C (summer), 5°C (winter) or 25°C for 2 months plus 2 months at 5°C (summer followed by winter–SW), and then moved to the germination temperatures. Embryo measurements were taken during pre-treatments and germination. Experimental seed burials were carried out for two populations of each species. Both species dispersed in summer. The population of A. nugorensis occurring on rocky outcrops differed in phenology from both the other A. nugorensis population from riparian vegetation and from A. barbaricina. Both species showed morphophysiological seed dormancy, with <50% germination under laboratory conditions. All riparian populations germinated only after the SW pre-treatment, while the rupicolous population germinated at 25°C, without any pre-treatment. Low germination percentages were observed in the experimental seed burials, suggesting the ability for both species to form a persistent soil seed bank.     

Germination traits and seed-bank dynamics of a biennial plant,Oenothera glazioviana Micheli

A study was conducted on the germination traits and seed-bank dynamics ofOenothera glazioviana (=O. erythrosepala), which sets seed in August in sand-dune systems in Japan. More than 90% of freshly matured seeds germinated over a wide range of temperature in light, but less than 10% did so in continuous darkness. Stratification (chilling under moist conditions) was ineffective in diminishing the light-requirement for germination. When fresh seeds were imbibed for 24 h including a 12-h light period, followed by 7-day air-drying, 94% of them became germinable in the dark at 25°C, but remained dormant at less than 15°C. of seeds collected in March from capsules of dead plants, 58% germinated in the dark at 25°C. After four cycles of alternatc 1-day wetting followed by 2-day drying or 1.5-day wetting followed by 1.5-day drying under a 12-h photoperiod, the fraction of viable seeds declined from 76% to 40% and 22%, respectively, due to germination during the wet periods. Seed-bag experiments were conducted in the field, using seeds given and not given a light-stimulus. Forty percent of the light-stimulated seeds germinated in the soil, whereas the seeds without a light-stimulus remained dormant throughout the experiment. When seeds were placed on the soil surface or at a depth of 0.5-1 cm, the proportion of germinable seeds declined during late spring and autumn, but not during winter and early spring. The seed-bank size of a natural population just prior to current seed dispersal was 2–3% of the seed production in the previous year, suggesting a high turnover rate of the seed-bank.     

Seed bank persistence of clonal weeds in contrasting habitats: implications for control

The ability of weeds to form a seed bank is important for their population dynamics and management because it provides a refuge enabling reinvasion after established target plants have died. However, knowledge of the differential seed behaviour of individual species over multiple years and varying environmental conditions is surprisingly rare but necessary for effective control of diverse weed populations. We established a seed burial experiment in alpine habitats differing in management regime (i.e., forest, hay meadow and pasture) to determine whether seeds of the unpalatable perennial weeds, Veratrum album (white hellebore) and Gentiana lutea (yellow gentian) were able to delay germination and remain viable over 3 years. Our study shows that both species formed a short-term persistent seed bank; in the third-year, the soil seed banks of both species were nearly depleted, having declined to <5% of their original size. Both species had strikingly different germination strategies: G. lutea seeds mainly germinated in their first-year, whilst the majority of V. album seeds germinated in their second-year. The fraction of dormant G. lutea seeds increased with seed age, indicating that seeds remained viable after forgoing germination in the previous year. Habitat-specific differences in seed germination increased with seed age, with germination fractions being lowest in moist hay meadows. This suggests that the negative effects of anoxic conditions became more pronounced as seeds aged in hay meadows. Conversely, seed dormancy was equal among habitats. The absence of a long-term persistent seed bank has important implications for the management of both nuisance and endangered-plant populations. In the case of V. album and G. lutea, re-colonization of habitats from the seed bank is unlikely after established plants have been removed.     

A Race for Survival: Can Bromus tectorum Seeds Escape Pyrenophora semeniperda-caused Mortality by Germinating Quickly?

BACKGROUND AND AIMS: Pathogen-seed interactions may involve a race for seed resources, so that seeds that germinate more quickly, mobilizing reserves, will be more likely to escape seed death than slow-germinating seeds. This race-for-survival hypothesis was tested for the North American seed pathogen Pyrenophora semeniperda on seeds of the annual grass Bromus tectorum, an invasive plant in North America. In this species, the seed germination rate varies as a function of dormancy status; dormant seeds germinate slowly if at all, whereas non-dormant seeds germinate quickly. METHODS: Three experimental approaches were utilized: (a) artificial inoculations of mature seeds that varied in primary dormancy status and wounding treatment; (b) naturally inoculated undispersed seeds that varied in primary dormancy status; and (c) naturally inoculated seeds from the carry-over seed bank that varied in degree of secondary dormancy, habitat of origin and seed age. KEY RESULTS: In all three approaches, seeds that germinated slowly were usually killed by the pathogen, whereas seeds that germinated quickly frequently escaped. Pyrenophora semeniperda reduced B. tectorum seed banks. Populations in drier habitats sustained 50 times more seed mortality than a population in a mesic habitat. Older carry-over seeds experienced 30 % more mortality than younger seeds. CONCLUSIONS: Given the dramatic levels of seed death and the ability of this pathogen to reduce seed carry-over, it is intriguing to consider whether P. semeniperda could be used to control B. tectorum through direct reduction of its seed bank.     

Variation in Predispersal Losses in Reproductive Potential in Rowan (<Emphasis Type="Italic">Sorbus aucuparia</Emphasis> L. Rosaceae) in the NW Iberian Peninsula

We identified the principal factors causing losses in reproductive potential (i.e. quantity of potentially dispersable seeds) over the predispersal period of the reproductive cycle of Sorbus aucuparia L. (Rosaceae) in various populations in the NW Iberian Peninsula. Over the period 1999–2002 we examined the magnitude and variability of losses observed in each stage from bud production to seed production (i.e. bud to flower, flower to initiated fruit, initiated to full-size fruit, and full-size to mature fruit). We also investigated spatial and temporal variation in these losses. Reproductive potential showed very significant losses over the predispersal period in all 4 years: in no case did the number of mature fruits available for dispersal exceed 20% of the initial number of flower buds. In all 4 years of study, we detected statistically significant variation among populations in total losses. The principal cause of loss of reproductive potential over the predispersal period in our Sorbus aucuparia populations was non-initiation of fruit. The plant’s interactions with predispersal predators of buds and seeds show high spatial and temporal variability. Also, the proportions of loss due to the different components did not show within-individual consistency over time. It thus seems unlikely that the different factors contributing to predispersal losses in rowan are currently generating significant selective pressures in our study populations.     

Fate of the soil seed bank of giant ragweed and its significance in preventing and controlling its invasion in grasslands

Giant ragweed (Ambrosia trifida, L. henceforth referred to as GR), an annual non‐native invasive weed, may cause health problems and can reduce agricultural productivity. Chemical control of GR in grasslands may have irreversible side effects on herbs and livestock. In an attempt to propose a solution to the harmful effects of GR on grasslands, this study explores the fate of its soil seed bank (SSB) and considers the physical control of its SSB reduction. By studying GR distributed in grasslands of the Yili Valley, Xinjiang, China, we measured the spatial and temporal changes in seed density, seed germination, dormancy, and death. We analyzed seed germination, dormancy, and death following different storage periods. The study analyzed population characteristics over time, including seed fate, and examined physical control methods for reducing the SSB density. The SSB of GR occurs in the upper 0–15 cm of soil in grasslands. Seed density in the SSB decreased by 68.1% to 82.01% from the reproductive growth period to the senescence period. More than 98.7% of the seeds were rotten, eaten, germinated, dispersed, or died within one year after being produced. The seed germination rate of the SSB decreased with the number of years after invasion. When stored for 0.5 or 3.5 years, seed germination rates fell by 40%, during which time seed death rate increased by almost 40%. When GR was completely eradicated for two consecutive years, the SSB and population densities decreased by >99%. The vast majority of GR seeds germinated or died within one year; the germination rate decreased significantly if the seeds were stored dry at room temperature for a long time. Newly produced seeds are the main source of seeds in the SSB. Therefore, thoroughly eradicating GR plants for several years before the seeds can mature provides an effective control method in grasslands.     

Effects of seed colour heterogeneity on germination behaviour of the desert plant Lotononis platycarpa (Fabaceae)

We aimed to determine the ecological role of three seed morphs observed for the first time in a desert population of Lotononis platycarpa (Fabaceae), with respect to their germination requirements. Seeds sorted by seed coat colour (olive green, orange and brown) were germinated under laboratory conditions under two photoperiods (12/12‐h light and continuous dark) and three alternating temperature regimes (15/25, 20/30, 25/35°C). We found that the three distinct seed types differ in their seed mass, germination percentage and speed of germination. Overall, the light‐incubated seeds germinated with higher percentages than seeds in the total darkness. Furthermore, seeds with orange coat germinated with higher percentages at 15/25 and 25/35°C (up to 60%, for both) and significantly faster than the other two colour morphs. Our results suggest an adaptive significance of seed colour heterogeneity in the harsh desert habitat inhabited by the study species.     

Population biology of Senecio integrifolius (Compositae), a rare plant in Sweden

In a study of natural populations of Senecio integrifolius in southern Sweden (1979–1983) seeds were found to disperse from mid-June to late July and most of them germinated in autumn. No soil seed bank was observed. On a heavily grazed site few seeds were produced but the percentage of estimated germination was high (c. 75%). In two moderately grazed habitats 8–10% of the estimated number of seeds produced in permanent plots germinated. In a lightly grazed habitat many seeds were produced but only a small percentage germinated (1.4%). In field experiments the average germination was 50–53% when newly harvested seeds were sown where the vegetation had been removed, and 20–33% when sown in undisturbed vegetation at a moderately grazed site. Germination was much lower in a lightly grazed habitat (3–12%). Survival of seedlings was much higher in heavily and moderately grazed habitats than in lightly grazed habitats. Mortality tended to be higher during the growing season (mid-April to early November) than during the winter, and increased markedly during a drought period in the summer of 1982. The half-life of plants established in 1980 varied from 39.3 years at the most intensively grazed site to 7.2 at the lightly grazed site. The number of flowering stems varied between years mainly according to weather. Few plants in the permanent plots flowered every year, the flowering being most frequent at heavily grazed sites. It is concluded that heavy grazing by cattle after seed dispersal is the appropriate management for maintaining S. integrifolius in Sweden.     

GEOGRAPHIC DIFFERENCES IN THE SEED GERMINATION OF PAPER BIRCH (BETULA PAPYRIFERA)

The effects of light and temperature on germination of paper birch (Betula papyrifera Marsh.) were examined in seeds from different geographic locations. Under continuous light at 14–18 C northern seed sources germinated more rapidly and achieved higher germination percentages than southern ones. Seed source influenced the range of temperatures over which germination occurred. Northern seeds had a wider temperature range than southern ones due to depression of the low temperature limit in northern seeds. At 15 C, seeds exhibited a photoperiodic-like response with progressive increases in germination as daily photoperiod was increased. This was not a true photoperiodic effect because seeds responded to the total amount of light rather than to the relative length of light and dark periods. It is suggested that at temperatures near 15 C, this response may be important in preventing fall germination. Prechilling at 3 C promoted subsequent dark germination in most seed sources tested, and northern seeds were much more sensitive to the treatments than southern seeds. In some seeds the prechilling stimulus was lost with protracted storage at 3 C; this was observed in both northern and southern seeds. Prechilling enhanced the seeds’ sensitivity to subsequent light. Although acquisition of the prechilling stimulus does not serve to measure the length of winter in these seeds, the high light sensitivity induced by prechilling probably accelerates spring germination. Northern seeds typically had thinner, more translucent pericarps than their southern counterparts.     

Influence of post-flowering temperature on seed development, and subsequent performance of crisp lettuce

Crisp lettuce plants cv. Saladin were grown from the time they started flowering, at 20/10°C (16 h day, 8 h night), 25/15°C and 30/20°C in glasshouses on two occasions in 1985. Yields of seed increased from, on average, 15 g to 27 g and then fell to 20 g per plant with progressive increases in temperature. The number of mature florets per plant increased with temperature but the number of seeds per mature floret was lower at 20/10°C and 30/20°C than at 25/15°C. An increase in temperature reduced mean seed weight by up to 45%, seed volume by 15%, cell numerical volume density (N_v) by 27% and the number of cells per seed by 39%. Percentage seed germination reached a maximum early in seed development at the stage when the pappus appeared through the involucral bracts. Differences in percentage germination and vigour of seeds (slope test) from different temperatures were accounted for largely by the effects on mean seed weight. However, when germinated at 30°C seeds produced at 30/20°C germinated more readily than those produced at 25/15°C or 20/10°C. Seed vigour gradually increased with an increase in the length of storage after harvest, reaching a maximum after 260 days. In general, seeds produced at 25/15°C exhibited a greater variation in numbers of seeds per floret, N_v, seed weight, times of seedling emergence, seedling and mature head weight than seeds produced at lower or higher temperatures.     

Germination ecology of the weed Parthenium hysterophorus in eastern Ethiopia

Germination ecology of Parthenium hysterophorus, recently introduced to Ethiopia, was studied in a series of experiments. Viability of the seeds was greater than 50% after 26 months of burial in the soil indicating the potential build‐up of a substantial persistent soil seed bank. A short period of dry storage was sufficient to overcome a light requirement for germination in a minor fraction of the seeds. Likewise, seeds exhumed from burial showed an increase in germination ability in darkness over time, with a weak tendency for seasonal cyclicity in dormancy level at one of two sites. Germination occurred at the mean minimum (10°C) and maximum (25°C) temperatures of the collection sites, as well as over a wide range of fluctuating (12/2°C‐35/25°C) temperatures in light. No germination of P. hysterophorus seeds occurred at osmotic potentials < ‐0.52 MPa (at 27°C), the species being less tolerant to moisture stress than sorghum grains. Most seedlings emerged from shallowly buried (< 0.5 cm) seeds and none from more than 5 cm depth. Naturally dispersed seeds required about 60 days, at a hot lowland site, to start emergence despite the presence of adequate rainfall, and higher number of seedlings emerged in undisturbed plots than in hand hoed plots. These experiments and field observations suggest that there are no obvious climatic conditions that may limit the germination of Parthenium hysterophorus in Ethiopia, but a high moisture requirement of the seeds for germination could be the major factor limiting germination during the dry season.     

Morpho‐colorimetric analysis and seed germination of Brassica insularis Moris (Brassicaceae) populations

Brassica insularis is a perennial plant growing on both coastal and inland cliffs. Three seed lots from Sardinia were analysed using an image analysis system to detect differences in seed morphology, both within and among populations. Germination requirements at constant (5–25 °C) and alternating temperatures (25/10 °C), both in light and in darkness, were evaluated for all populations. In addition, the effect of a dry after‐ripening period (90 days at 25 °C) was also investigated. Morpho‐colorimetric analysis clearly identified seeds from different populations and discriminated three chromatic categories for seeds belonging to the Isola dei Cavoli coastal population, but not for the inland Masùa and the coastal Planu Sartu. Inter‐population variability was also observed in germination behaviour. B. insularis seeds germinated, with percentages up to 60%, in a wide range of temperatures (5–25 °C), and neither light nor dry after‐ripening affected final germination percentages. Moisture content measurements were made for seeds of each colour, but there were no particular differences among colours. Inter‐populational variability in germination behaviour may be a survival strategy for species growing under unpredictable environmental conditions, such as under Mediterranean climate, while heteromorphy may be due to independent evolutionary divergence processes of the Isola dei Cavoli population.     

Elevated CO2 decreases seed germination in Arabidopsis thaliana

The impact of elevated [CO₂] on seed germination was studied in different genotypes of Arabidopsis thaliana from natural populations. Two generations of seeds were studied: the maternal generation was produced in the greenhouse (present-day conditions), the offspring generation was produced in two chambers where the CO₂ concentration was either the present atmospheric concentration (about 350 ppm) or elevated (700 ppm). The seeds were tested for proportion of germinated seeds and mean germination time in both chambers to study the impact of elevated [CO₂] during seed production and germination. Elevated [CO₂] during maturation of seeds on the mother-plants decreased the proportion of germinated seeds, while elevated [CO₂] during germination had no effect on the proportion of germinated seeds. However, when seeds were both produced and germinated under elevated [CO₂] (situation expected by the end of next century), germination was slow and low. Moreover, the effect of the [CO₂] treatment differs among genotypes of Arabidopsis: there is a strong treatment × genotype interaction. This means that there is ample genetic variance for a selective response modiying the effects of high levels of [CO₂] in natural populations of Arabidopsis thaliana. The outcome at the community level will depend on what seeds are available, when they germinate and the resulting competition following germination.     

Experimental determination of seed emergence and carry‐over in the soil seed bank of the herbaceous perennial,Trachymene incisa (Apiaceae)

Abstract Many populations of herbaceous perennial plants contain seeds stored in a soil seed bank. The contribution of seeds to population persistence is an important parameter in population models but germination rates of known‐age seeds are difficult to obtain because individual seeds cannot easily be followed. Although Trachymene incisa Rudge plants produce copious seeds that are dispersed into the soil, the existence of a seed bank has not been confirmed. To quantify the potential for a seed bank fresh seeds of T. incisa were sown into experimental seed banks in the eucalypt‐dominated Agnes Banks Woodland in western Sydney, NSW. A recent fire provided the opportunity to compare germination in the burnt and unburnt vegetation. Density of seed sowing and time of maturation/dispersal of seeds were manipulated in 75 seed cages. Emergence of seeds after 5 months was significantly higher for the earliest planting date but after 1 year, germination of seeds planted in the later weeks increased, and the final germination for all weeks was 28%. Density of sowing and the recent fire did not affect emergence. A second experiment planted over a broader time span (9 weeks instead of 3 weeks) confirmed the effect of planting date but also found significant spatial variation on a scale of tens of metres. Laboratory germination rates of over 70% confirmed that the seeds were viable and non‐dormant when sown in the field cages. The carry‐over of non‐germinated seed in the soil seed bank is estimated to be about 70% after 2 years, implying that a cohort of seeds would not be depleted through germination alone for up to 40 years. The potential for a long‐lived seed bank in this species is interesting because the plants are also capable of resprouting from their rootstock after fire, giving them characteristics of both resprouters and seeders.     

Germination ecology of <Emphasis Type="Italic">Scorpiurus subvillosus</Emphasis> L. seeds: the role of temperature and storage time

Scorpiurus subvillosus L., wide spread in pastures of Mediterranean basin, is disappearing in the native pastures of the Hyblean plateau (Sicily, southern Italy), because of overgrazing and intensive management techniques. Moreover, it exhibits seed coat dormancy, which delays and reduces germination preventing its diffusion. This paper represents a first attempt in order to investigate changing in germination determined by storage time and temperature on seeds of two populations of S. subvillosus. Germination of S.␣subvillosus seeds was tested in relation to four storage time (30, 130, 200 and 360 days after harvest (DAH)), eight constant temperatures (5, 10, 15, 20, 25, 30, 35 and 40°C) and two populations of different provenience (30 and 600 m above mean sea level). The experiments were conducted either on scarified and unscarified seeds. In S. subvillosus the failure of germination under favourable conditions must be attributed␣only to seed coat, since seed scarification enhanced germination percentage with values up to 100% at almost all tested temperatures. In both treatments, but with a grater incidence in unscarified, seed germination increased gradually as temperature raised, peaking at 20–25°C, then declined with further increases of temperatures. At 40°C no germination occurred. Storage time induced a softening effect, which is somewhat limited by the natural ageing of seeds occurring from about 6 months after harvest.     

Effects of cold storage of germinated vegetable seeds prior to fluid drilling on emergence and yield of field crops

Germinated seeds of five vegetable crops were stored at 1 °C in aerated water or humid air for periods of up to 15 days and then fluid drilled in the field. Viability of germinated seed was maintained in carrot and parsnip for a minimum of 15 days, salad onion for 9 days, lettuce for 6 days and cabbage for 3 days. When salad onion, lettuce and cabbage seeds were stored for longer than these times there was a reduction in the number of seedlings emerging. In species where storing seeds for up to 15 days caused a reduction in emergence an adjustment to the seed rate related to the length of storage can be made to achieve the desired population and still retain the advantages of fluid drilling pre-germinated seeds. This is possible because with one exception (an increase in the variation of salad onion stem diameters) the performance of plants from all stored germinated seeds after emergence was similar to plants from freshly germinated seeds.     

Seed size variation and predation of seeds produced by wild and crop-wild sunflowers

The movement of pollen between crop and wild sunflowers (both Helianthus annuus) has led to concerns about the possible introduction of crop transgenes into wild populations. The persistence of crop traits in wild populations will depend in part on the relative fitness of crop-wild hybrid vs. wild plants. Using seeds from two large experimental field plots, we found that seeds produced by crop-wild plants were twice the size of wild seeds and differed in coloration. Head diameter, date of flowering, identity of mother plant, and levels of predispersal predation explained some variation in mean seed size. We hypothesized that postdispersal vertebrate seed predation would be affected by seed size, with hybrid seeds preferentially eaten. In each of three field trials, significantly more hybrid seeds were eaten (62% of hybrid seed; 42% of wild seed). Within the category of wild seeds, larger seeds were preferentially eaten; however among hybrid seeds, predation was not significantly related to seed size. In this study, differential predation thus reduces hybrid fitness and would presumably slow the spread of transgenes into wild populations.