Germination of Ficus insipida (Moraceae) seeds from toucan (Ramphastos sulfuratus) and spider monkey (Ateles geoffroyi) feces

To test the null hypothesis that two vertebrate fructivores, toucans (Ramphastos sulfuratus) and spider monkeys (Ateles geoffroyi), are equally specialized in germinating Ficus insipida seeds after these have passed through their digestive tracts, we fed fruits to captive individuals. We extracted seeds from feces and placed them on filter paper in petri dishes under controled light, temperature and humidity. Control seeds had not passed through a digestive tract. We found that a greater proportion of seeds from A. geofroyi (65%) germinated (R. sulfuratus: 4%). The germinative value was also greater in seeds from monkey feces (rate=13.76; toucan 0.046; control group 0.172). If, despite individual variability of seeds and dispersers, future studies continue to show that A. geoffroyi favors germination more (maybe because of a longer digestion time), this would indicate that diet specialization is not necessarily related to dispersal efficiency.     

Nondeep simple morphophysiological dormancy in seeds of Ilex maximowicziana from northern (subtropical) and southern (tropical) Taiwan

In this study, we show that seeds of Ilex maximowicziana collected from northern and southern Taiwan differ in germination responses to temperature. Seeds produced by plants growing in the tropical environment of southern Taiwan were more responsive (in a positive way) to higher incubation temperatures than those produced by plants growing in the subtropical environment of northern Taiwan. On the other hand, seeds produced in northern Taiwan were more responsive (in a positive way) to low incubation temperatures and to cold stratification than those from southern Taiwan. Germination percentages and rates of seeds from northern Taiwan were higher at 20/10°C than at 30/20°C, reaching a plateau of >80% germination after 12 weeks incubation, whereas germination of seeds from southern Taiwan reached >80% at 30/20 and 25°C but not at 20/10°C. Gibberellic acid (GA₃) increased germination rate but not germination percentage of seeds from both southern and northern Taiwan. Freshly matured seeds of I. maximowicziana have rudimentary embryos. During dormancy break, embryo length increased 11.5- and 8.0-fold in northern and southern seeds, respectively, before radicle emergence. Thus, seeds of Ilex maximowicziana have nondeep simple morphophysiological dormancy. This is the first detailed study of the germination requirements of a subtropical/tropical species of the large cosmopolitan genus Ilex.     

The nutritional value of narrow-leafed lupine (Lupinus angustifolius) for fattening pigs

The aim of this study was to determine the nutrient digestibility of seeds of four varieties of narrow-leafed lupines (Lupinus angustifolius) and the possibility of soya bean meal (SBM) substitution by lupine seeds alone and in combination with rapeseed meal (RSM) in the diets of pigs. The seeds of the lupine varieties Kalif, Sonet, Zeus and Boruta were analysed. The apparent total tract digestibility (ATTD) was determined on 50 cross-bred pigs using the difference method with titanium dioxide as a marker. The substitution of SBM by lupine seeds alone (at 0 – 100%) was tested on 60 pigs (20–105 kg body weight (BW)) and by a combination of lupine seeds and RSM on 180 fattening pigs (35–80 kg BW). The chemical composition of lupine seeds differed considerably, especially in terms of crude protein and mineral content. All seeds contained less than 0.05% alkaloids and 9.3% oligosaccharides in dry matter. The ATTD of protein ranged from 70% to 74%, those of ether extract from 36% to 55% and those of gross energy from 77% to 84%. The entire replacement of SBM by lupine seeds (var. Sonet) did not have a negative effect on the performance of grower and fattener pigs. The substitution of SBM by a combination of lupines and RSM reduced the performance of growing and finishing pigs significantly.     

Vicilin variants and the resistance of cowpea (Vigna unguiculata) seeds to the cowpea weevil (Callosobruchus maculatus)

1. A globulin fraction prepared from the meal of Callosobruchus maculatus-resistant cowpea (Vigna wiguiculata) seeds was shown to be detrimental to this bruchid when incorporated in artificial seeds.2. The performance of C. maculatus was also shown to be strongly hindered by vicilins from resistant seeds when these storage proteins were incorporated in artificial seeds at the level of 2%.3. The purified vicilins from seeds of both resistant and susceptible cowpea varieties were shown to have the same SDS-PAGE pattern but different mobilities in non-denaturing polyacrylamide gel electrophoresis.4. These results and previous ones obtained by us (Silva and Xavier-Filho, 1991; Sales et al., 1992) strongly suggest that the resistance of cowpea seeds from the cultivar TVu 2027 and from others bred from it is associated with the presence of vicilin molecules which are refractory to digestion by bruchid midgut proteinases.     

POLLEN-MEDIATED GENE FLOW IN CUCURBITA FOETIDISSIMA (CUCURBITACEAE)

Pollen-mediated gene flow along a linear series of patches of the functionally gynodioecious, bee-pollinated Cucurbita foetidissima was assessed using electrophoretic analysis of seed allozymes. Gene flow was documented at distances up to 0.7 km. For the 13 patches examined, interpatch pollen must have sired from 0% to 48.3% (X̄ = 8.5%) of the seeds of monoecious plants (hermaphrodites). Rates of interpatch siring of seeds of pistillate plants (females) averaged 20.4% (range 8.6%-40%) for the three patches examined. Heterogeneity among fruits in seeds sired by interpatch pollen indicates that the arrival of interpatch pollen is clumped with respect to stigmas. Within patches, plants of the same sex type usually shared identical five-locus genotypes, suggesting that clonal propagation predominates. Since approximately 90% of seeds are sired by intrapatch pollen, seeds of monoecious plants appear to result primarily from geitonogamous (self-) fertilization. This may help explain the existence of female plants in natural populations, since self-fertilization has been shown to severely reduce the survival of seedlings in this species.     

NONDEEP COMPLEX MORPHOPHYSIOLOGICAL DORMANCY IN SEEDS OF OSMORHIZA CLAYTONII (APIACEAE)

Freshly matured seeds of Osmorhiza claytonii exhibit a type of morphophysiological dormancy (MPD). Under natural conditions, embryo growth begins in late September and early October and continues until mid***- to late February, with the peak in October and November. Most seeds germinate between mid-February and late March. Embryos did not grow in seeds incubated for 24 weeks at 30/15 (warm stratification) or 5 C (cold stratification). However, in seeds given 12 weeks at 30/15 and then 12 weeks at 5 C, embryo length increased 1,246% while seeds were at 5 C. Zero to 7 days of warm followed by 24 weeks of cold stratification resulted in 2%–27% germination of fresh seeds, whereas 2 to 12 weeks of warm followed by 24 weeks of cold stratification resulted in 80%–98% germination. Warm plus cold stratification was required for embryo growth and germination of seeds that remained undispersed for a year in the field. GA₃ was partially effective in substituting for warm stratification. The name “nondeep complex MPD” is proposed for the type of MPD found in O. claytonii and a few other species, making a total of eight types of MPD presently known.     

Seed predation and dispersal of glabrous filbert (Corylus Heterophylla) and pilose filbert (Corylus Mandshurica) by small mammals in a temperate forest, northeast China

We investigated the seed dispersal of glabrous filbert (Corylus heterophylla) and pilose filbert (Corylus mandshurica), two large-seeded shrub species in a temperate forest, northeast China, September 2006. Small mammals such as Apodemus speciosus, Clethrlonomys rufocanus, and Eutamias sibiricus, were regarded as the main dispersal agents. More seeds were harvested by small mammals in pilose filbert (98%) than in glabrous filbert (87.5%) till our last survey. Seed removal rates differed between the two species. Fewer seeds of glabrous filbert (17.5%) were eaten in situ than pilose filbert (57.5%). More seeds of glabrous filbert were removed (70%), stay intact after removal (25.5%), eaten after removal (16%) than pilose filbert. However, more seeds were cached after removal in pilose filbert than in glabrous filbert (10.5 and 4%, respectively). Fewer tagged seeds of pilose filberts (14%) were missed than glabrous filberts (24.5%). About 8 and 12 primary caches were found in glabrous filbert and pilose filbert seeds respectively, indicating scatter hoarding. All of the removed seeds were distributed within 10 m of seed stations for both filberts. The average dispersal distances for glabrous filbert did not differ from pilose filbert. Only a small proportion of the caches remained till our last survey (2 and 1%, respectively). Based on the results, we found a difference in dispersal patterns of glabrous filbert and pilose filbert seeds. Evidences showed that glabrous filberts might be a less preferred seed species for small seed-eating mammals compared with pilose filbert, probably due to its harder and thicker husk and low seed profitability.     

REPRODUCTIVE ECOLOGY OF JEFFERSONIA DIPHYLLA (BERBERIDACEAE)

The reproductive ecology of Jeffersonia diphylla (L.) Pers. (Berberidaceae) was investigated by studying its breeding system, ovule production, seed set, seed dispersal by ants and seed predation by rodents. This species flowers early in the spring and is facultatively autogamous. In a typical year fruit and seed set is high (90%), however, freezing temperatures from late spring frosts in 1983 and 1985 resulted in low fruit set (7% and 20%, respectively), and reduced seed set in those flowers that produced fruit. No differences in seed set between selfed and outcrossed flowers were observed over a two-yr period (1983–84). Ovule number per capsule increased with plant size as measured by leaf number. Seed set and seed wt were unaffected by leaf number unless leaves were removed after flowering was initiated. Jeffersonia diphylla is myrmecochorous. Ants removed seeds faster when seeds were placed in areas where J. diphylla plants were absent, suggesting that dispersal within J. diphylla populations is ant limited. Moreover, fresh (1 day old) seeds were removed by ants faster than 3 day old seeds. Seed predation by rodents prior to dehiscence from capsules is heavy in large populations (85–90%), and apparently negligible in small populations. Predation of seeds that are released from capsules is heavy (approx. 66%), particularly at night. Overall, seed predators consume about 96% of the seed crop in well established populations, but probably much less in small young populations. Hence, seedling recruitment is likely to be higher in small populations, whereas ramet production from rhizomes is the primary mode of propagation in large ones. The evolution of autogamy, early flowering, and myrmecochory are discussed in light of the results of this study.     

Evaluation of a mechanical seed planter for transplanting Zostera marina (eelgrass) seeds

Few seagrass transplant projects worldwide have relied on seeds, and those projects using Zostera marina (eelgrass) seeds have generally found low rates of seedling establishment (<10%). We compared seedling establishment achieved by a mechanical seed planter with seeds broadcast on the sediment surface by hand. The planter injected seeds into the sediment by pumping the seeds, suspended in a gelatin-based matrix, to a benthic sled with eight planting nozzles. As a control for the gel, seeds were also injected into the sediment without gel using a hand-held pipette. Seeds were planted at a density of 300 m⁻² into six replicate plots at each of three sites in the Chesapeake Bay region in September 2005, with seedling establishment measured in April 2006. Burying seeds, either with or without gel, had an overall positive effect on seedling establishment, but the effectiveness and the best method varied among sites. Mean seedling establishment for machine-planted seeds was significantly greater than for broadcast seeds at the Piankatank River site (4% vs. 1%), but not at the York (1.2% vs. 1.4%) or Spider Crab Bay (10.1% vs. 7.4%) sites. The effect of the gel was inconsistent among sites, with the highest seedling establishment (18.8%), resulting from seeds injected by pipette without gel at the Spider Crab site. Seed burial shows promise for increasing seedling establishment relative to seed broadcasting in the Chesapeake region, but further investigation of seed–sediment interactions at specific restoration sites is necessary. Low seedling establishment rates remain a bottleneck for seed-based eelgrass restoration.     

Foraging and pollination behaviour of Xylocopa calens Lepeletier (Hymenoptera: Apidae) on Phaseolus coccineus L. (Fabaceae) flowers at Yaounde (Cameroon)

To evaluate impact of carpenter bee, Xylocopa calens, on pod and seed set of Phaseolus coccineus, its foraging and pollinating activities were studied in Yaounde, for two seasons (May–July 2008 and April–June 2009). Observations were made on 40 inflorescences per treatment. The treatments included unlimited floral access by all visitors, bagged flowers to deny all visits, and limited visits by X. calens only. In addition, all flower visitors were recorded. The carpenter bee's seasonal rhythm of activity, its foraging behavior on flowers, its pollination efficiency, the fructification rate and the number of seeds per pod were recorded. Individuals from 16 species of insects were recorded visiting flowers of P. coccineus in the 2 years. Xylocopa calens was the most frequent, followed by Chalicodoma cincta cincta and Apis mellifera. Apart from bees, wasps were also recorded as likely predators. Xylocopa calens mainly foraged for nectar resources. The mean foraging speed was 9.62 flowers/min. Flowers visited by X. calens had higher fruiting rate compared with others, while those bagged had the lowest. In addition, seed formation was higher in X. calens‐visited flowers compared with all others. The results show that this crop experiences pollination deficit even under normal circumstances, considering that flowers visited by X. calens had higher yields compared with those under unlimited access by all visitors. The fruiting rate, the number of seeds/pod and the percentage of normal seeds of unprotected inflorescences were significantly higher than those of inflorescences protected from insects. X. calens foraging resulted in a significant increment of the fruiting rate by 25.80%, as well as the number of seeds/pod by 14.97% and the percentage of normal seeds by 27.75% in 2008 and 18.39% in 2009. Conservation of X. calens nests close to P. coccineus fields could be recommended to improve pod and seed production in the region.     

Red-Tailed Guenons (Cercopithecus ascanius) and Strychnos mitis: Evidence for Plant Benefits Beyond Seed Dispersal

I report data collected on red-tailed guenon (Cercopithecus ascanius schmidti) fruit processing behaviors between June 1993 and April 1994. Red-tailed guenons consumed the fruit of Strychnos mitis in 542 of 2,930 fruit-eating events (FEEs). The monkeys spat out cleaned seeds of Strychnos mitis in a majority of these records (477/542; 88%); seeds were occasionally swallowed whole, but only when pulp was unripe (69/542; 12%). In 83% of the FEEs on Strychnos, the red-tailed guenons spat out seeds within 10 m of the removal site; they typically stayed in the same tree while processing fruit, and in 56% of the FEEs, they moved <1 m before spitting seeds. I monitored spat seeds to evaluate the impact of monkey fruit processing on seed fate. Results indicate that 83% of seeds spat out by the red-tailed guenons germinated, while only 12% of unprocessed seeds survived to germination (p < 0.01). Of the processed seeds that germinated, 60% survived to germination and seedling establishment, while only 5% of unprocessed seeds survived to seedling establishment (p < 0.01). Unprocessed seeds were also more likely to be attacked by seed predators (p < 0.01) and fungus (p < 0.01). Although there is generally high mortality in seeds/seedlings, mature trees of Strychnos mitis are found in groves of adults, under which dense populations of seedlings and saplings can occur. These data suggest that Strychnos mitis does not conform to expectations of the Janzen-Connell model of seed escape from parent trees. Instead, I suggest that by removing pulp, a process that results in a reduction of fungal pathogen attack, red-tailed guenons positively effect the seed survivorship of Strychnos mitis. Although this effect has been observed in pulp-cleaning ant species, it is a hitherto undescribed effect of primates on their fruit resources.     

The effect of sowing pre-germinated seeds of lettuce (Lactuca sativa) on seedling emergence

Glasshouse and field experiments were conducted to investigate the effects of sowing both untreated and pre-germinated seeds, with or without a gel carrier, and also pelleted seeds into soils of different moisture contents and at different temperatures on seedling emergence of lettuce cvs Cobham Green and Avondefiance. Pre-germinated seeds with radicles 1–2 mm long gave, on average, 15 and 4% higher final percentage emergence than untreated and pelleted seeds respectively. Sowing pre-germinated seeds having radicles longer than 2 mm gave more variable levels of emergence and often reduced emergence, particularly when they were sown into dry or drying seedbeds. Increasing the rate of gel carrier from 0 to 30 ml m^_1 of row reduced emergence progressively from 67 to 30%. Pre-germinated seeds emerged 2–5 and 3-0 days earlier than untreated and pelleted seeds, respectively, at a soil temperature of 10 ^oC; at 15 ^oC the corresponding figures were 0–5 and 3–5 days. Sowing pre-germinated seeds reduced the spread (i.e. variability) of emergence by an average of 37 and 47% compared with untreated and pelleted seeds, respectively.     

Seed handling in chimpanzees (Pan troglodytes) and redtail monkeys (Cercopithecus ascanius): Implications for understanding hominoid and cercopithecine fruit-processing strategies and seed dispersal

Primates are confronted with an array of constraints in feeding on fruit, including the removal of adhesive, energy-rich pulp from seeds. In this paper, I discuss how primates meet this challenge and present data on the fruit-processing and seed-handling behavior of chimpanzees and redtail monkeys in Kibale National Park, Uganda. These data are then related to these species' services as seed dispersers. Particular attention was paid to the methods by which primates removed pulp from seeds, the density of seed clumps that they deposited (by spitting, dropping, or defecating) to the forest floor, and the distance seeds were moved from parent trees. Distance and density differences in chimpanzee and redtail seed dispersal resulted from distinct fruit-processing and seed-handling methods. It was observed, in general, that redtail monkeys engaged in fine oral processing and were seed spitters: most seeds were dispersed in close proximity to parent trees (84% of spat seeds <10 m of parent tree), and deposited singly (100% seeds spat singly). In contrast, chimpanzees were coarse fruit processors and seed swallowers: seeds were defecated in denser clumps (e.g., a mean of 149 large seeds/dung sample and hundreds of small seeds/dung sample), far from parent trees. I evaluate the factors that shape patterns of fruit processing in hominoids and cercopithecines, and argue that the observed seed handling differences can be attributed to differences in digestive retention times, oral anatomy, and alternative mechanisms by which to avoid the cost of seed ballast. Am J Phys Anthropol 109:365–386, 1999. © 1999 Wiley-Liss, Inc.     

Effects of black howler monkey (Alouatta caraya) seed ingestion on insect larvae

Fig (Ficus monckii) and laurel (Ocotea puberula) seeds were obtained from Alouatta caraya feces (ingested seeds) and from trees (noningested) in northeastern Argentina. Seeds were examined to detect the presence (infested) or absence (noninfested) of larvae. Sixty percent (N = 315) of noningested fig seeds were galls with insect larvae inside, while 23% (N = 331) were encountered in the ingested group. Eighty-two percent (N = 28) of noningested laurel seeds were infested, and only 19% (N = 63) of ingested seeds were infested. According to the present data, the insects' larvae are digested by howlers intaking animal protein, but the laurel seeds were not destroyed. Am. J. Primatol. 45:411–415, 1998. © 1998 Wiley-Liss, Inc.     

Allozyme evidence for polyzygotic polyembryony in Siberian stone pine (Pinus sibirica Du Tour)

Approximately 4000 mature seeds from 350 trees in nine populations (12–75 trees per population) of Siberian stone pine were investigated for multiple embryos (polyembryony). Haploid megagametophytes and embryos were genotyped for eight allozyme loci. Eight-yone seeds (2.11%) had more than 1 embryo. Of these, 71 seeds had 2 embryos (1.85%), 6 seeds had 3 embryos (0.16%), 3 seeds had 4 embryos (0.08%) and 1 seed had 6 embryos (0.026%). Allozyme comparison of megagametophytes and embryos could distinquish two types of polyembryony in 56 of the 81 seeds. In 28 seeds (50%) the polyembryony was polyzygotic (independent fertilizations of more than one egg cell in the ovule); 25 seeds (45%) had most likely monozygotic polyembryony (genetically identical embryos resulting from the cleavage of a single proembryo) and 3 seeds had both genetically different and genetically identical embryos. To the best of our knowledge, this is the first genetic evidence for the form of polyembryony in conifer seeds.     

Infection of Spathoglottis plicata (Orchidaceae) seeds by mycorrhizal fungus

Spathoglottis plicata seeds were encapsulated in 4-mm-diameter capsules of alginate-chitosan or alginate-gelatin and infected with the mycorrhizal fungus Rhizoctonia AM9. The encapsulated seeds were placed directly on Rhizoctonia culture. About 66% of the seeds encapsulated in sucrose-free chitosan-alginate established a symbiotic relationship with the mycorrhizal fungus after co-culturing for 2 weeks. The highest percentage of infection observed was about 84%. Addition of sucrose or using gelatin-alginate for encapsulation reduced the percentage of infection by about half. The growth of Rhizoctonia AM9 in sucrose-free alginate, chitosan and gelatin was found to be minimal. The advantages of germinating orchid seeds, encapsulated in sucrose-free polymers, through mycorrhizal infection is discussed. Received: 19 February 1998 / Revision received: 8 May 1998 / Accepted: 20 May 1998     

DORMANCY IN SEEDS OF FRASERA CAROLINIENSIS (GENTIAN ACEAE)

In freshly matured seeds of the long-lived monocarpic perennial, Frasera caroliniensis Walt., the embryos are underdeveloped and physiologically dormant. Dormancy was broken by a long period of stratification (chilling) at 5 C. Seventy six percent of the seeds germinated at 20 C (day)/10 C (night) after 98 days of chilling at 5 C, while seeds kept at 5 C germinated to 87% after 205 days. A warm, moist pretreatment was not required for subsequent breaking of dormancy at 5 C. Embryos in fresh seeds averaged 1.3 mm long, but after 12 weeks of chilling they averaged 4.1 mm. Thus, the embryos require a period of chilling to become fully developed, after which seeds can germinate at the afterripening temperatures (5 C) or at some higher temperature. Seeds of F. caroliniensis fit Nikolaeva's (1977) morpho-physiological complex dormancy type.     

Evidence of Eelgrass (Zostera marina) Seed Dispersal by Northern Diamondback Terrapin (Malaclemys terrapin terrapin) in Lower Chesapeake Bay

The initial discovery in May 2009 of eelgrass (Zostera marina) seeds in fecal samples of wild-caught northern diamondback terrapins (Malaclemys terrapin terrapin) was the first field evidence of eelgrass seed ingestion in this species. This finding suggested the potential of terrapins as seed dispersers in eelgrass beds, which we sampled for two additional years (2010 and 2011). Seeds were only found in feces of terrapins captured prior to June 8 in all three years, coinciding with eelgrass seed maturation and release. Numbers of seeds in terrapin feces varied annually and decreased greatly in 2011 after an eelgrass die off in late 2010. The condition of seeds in terrapin feces was viable-mature, germinated, damaged, or immature. Of terrapins captured during time of seed release, 97% were males and juvenile females, both of which had head widths <30 mm. The fraction of individuals with ingested seeds was 33% for males, 35% for small females, and only 6% for large (mature) females. Probability of seed ingestion decreased exponentially with increasing terrapin head width; only males and small females (head width <30 mm) were likely to be vectors of seed dispersal. The characteristic that diamondback terrapins have well-defined home ranges allowed us to estimate the number of terrapins potentially dispersing eelgrass seeds annually. In seagrass beds of the Goodwin Islands region (lower York River, Virginia), there were 559 to 799 terrapins, which could disperse between 1,341 and 1,677 eelgrass seeds annually. These would represent a small proportion of total seed production within a single seagrass bed. However, based on probable home range distances, terrapins can easily traverse eelgrass meadow boundaries, thereby dispersing seeds beyond the bed of origin. Given the relatively short dispersion distance of eelgrass seeds, the diamondback terrapin may be a major source of inter-bed seed dispersal and genetic diversity.     

Germination requirements inCarex kobomugi (Sea Isle)

Laboratory and field germination experiments inCarex kobomugi seeds were pursued to clarify their germination requirements and availability of the requirements in the field. In the laboratory experiments, more than 50% of the seeds ofC. kobomugi germinated under 35/30C or 25/20C when they were scarified with 98% H₂SO₄ after removal of their utricles, and chilled in moist condition for 28 to 42 d. Seeds with utricles or those without scarified with H₂SO₄ did not germinate. Seeds sown at 10-cm depth at the Kado-ori coast on 11 February 1991 after soaked in H₂SO₄ showed 40% germination by 29 April 1991, whereas those without H₂SO₄ treatment did not germinate. These results suggest that seed-coat impermeability and embryo immaturity are possible causes of the dormant state in seeds ofC. kobomugi ripen in summer. In the field, the moist-chilling condition is available in winter and the seeds can germinate in the following spring if the seed-coat impermeability is relaxed before winter.     

Effect of seed predation on seed bank size and seedling recruitment of bush lupine (Lupinus arboreus)

Whether seed consumers affect plant establishment is an important unresolved question in plant population biology. Seed consumption is ubiquitous; at issue is whether seedling recruitment is limited by safe-sites or seeds. If most seeds inhabit sites unsuitable for germination, post-dispersal seed consumption primarily removes seeds that would otherwise never contribute to the population and granivory has minimal impacts on plant abundance. Alternatively, if most seeds ultimately germinate before they lose viability, there is greater potential for seed consumption to affect plant recruitment. Of the many studies on seed consumption, few ask how seed loss affects seedling recruitment for species with long-lived seed banks. We examined post-dispersal seed predation and seedling emergence in bush lupine (Lupinus arboreus), a woody leguminous shrub of coastal grasslands and dunes in California. We followed the fate of seeds in paired experimental seed plots that were either protected or exposed to rodent granivores in grassland and dune habitats. Significantly more seeds were removed by rodents in dunes than grasslands. In dunes, where rodent granivory was greatest (65% and 86% of seeds removed from plots by rodents in two successive years), there is a sparse seed bank (6.6 seeds m⁻²), and granivory significantly reduced seedling emergence (in the same two years, 18% and 19.4% fewer seedlings emerged from exposed versus protected plots), suggesting seed rather than safe-site limited seedling recruitment. In contrast, rodents removed an average of 6% and 56% of seeds from grassland plots during the same two years, and the grassland seed bank is 43-fold that of the dunes (288 seeds m⁻²). Even high seed consumption in the second year of the study only marginally influenced recruitment because seeds that escaped predation remained dormant. Burial of seeds in both habitats significantly reduced the percentage of seeds removed by rodents. Results suggest that granivores exert strong but habitat-dependent effects on lupine seed survival and seedling emergence. Received: 24 October 1996 / Accepted: 4 February 1997