Effects of seed aggregation on the removal rates of elaiosome-bearing Chelidonium majus and Viola odourata seeds carried by Formica polyctena ants

In a field dominated by Formica polyctena Foerst. ants, we examined the effect of seed aggregation on the seed-removal rates of two plant species: a large-seeded obligate myrmecochore Viola odorata L. and a small-seeded diplochore Chelidonium majus L., which was autochorous as well as myrmecochorous. The effect was statistically non-significant in V. odorata but significant in C. majus, with more closely aggregated seeds having higher removal rates. The large seeds of the obligate myrmecochore were more quickly discovered and repeatedly removed by ant workers than were the small seeds of the diplochore.     

Seed handling behaviours of native and invasive seed‐dispersing ants differentially influence seedling emergence in an introduced plant

相似文献   

Indirect effects of granivory by harvester ants: plant species composition and reproductive increase near ant nests

Summary Of 36 plant species surveyed, 6 were significantly associated with nests of the desert seed-harvester ant Veromessor pergandei or Pogonomyrmex rugosus; two other plant species were significantly absent from ant nests. Seeds of two common desert annuals, Schismus arabicus and Plantago insularis, realize a 15.6 and 6.5 fold increase (respectively) in number of fruits or seeds produced per plant growing in ant nest refuse piles compared to nearby controls. Mass of individual S. arabicus seed produced by plants growing in refuse piles also increased significantly. Schismus arabicus, P. insularis and other plants associated with ant nests do not have seeds with obvious appendages attractive to ants. Dispersal and reproductive increase of such seeds may represent a relatively primitive form of ant-plant dispersal devoid of seed morphological specializations. Alternatively, evolution of specialized seed structures for dispersal may be precluded by the assemblage of North American seed-harvester ants whose workers are significantly larger than those ants normally associated with elaiosome-attached seed dispersal. Large worker size may permit consumption of elaiosome and seed.     

蚂蚁觅食与搬运行为对阜平黄堇和小花黄堇种子散布的影响

蚂蚁是无脊椎动物中重要的种子传播者,蚂蚁散布影响植物种子的传播和扩散,进而会影响种苗的空间分布格局。在野外研究了蚂蚁觅食及搬运行为对阜平黄堇(Corydalis wilfordii Regel)和小花黄堇(C.racemosa(Thunb.)Pers.)种子散布的影响。结果显示,双针棱胸蚁和束胸平结蚁是两种植物种子的共同搬运者,前者行使群体募集,后者行使简单协作募集。在搬运阜平黄堇种子时,双针棱胸蚁在原地或搬运途中取食油质体后抛弃的种子约占种子总数的56%,而拖至蚁巢的种子约占种子总数的44%,平均搬运距离为(1.85±0.24)m,搬运效率为(43.8±7.5)粒/h;而束胸平结蚁将完整种子全部直接搬运至蚁巢,平均搬运距离为0.45 m,搬运效率为(7.3±2.2)粒/h。在搬运小花黄堇种子时,双针棱胸蚁和束胸平结蚁均将完整种子全部直接搬运至蚁巢,平均搬运距离分别为(6.27±4.40)m和(6.65±1.64)m,搬运效率分别为(34.2±6.5)粒/h和(10.6±3.2)粒/h。这说明行使群体募集的蚂蚁比行使简单协作募集的蚂蚁有较高的搬运效率,蚂蚁散布导致阜平黄堇和小花黄堇种子到达蚁巢的数量和搬运距离不同,而这种不同与相应搬运蚂蚁的觅食对策、搬运行为和种子特征有关。阜平黄堇种子比小花黄堇种子大,但阜平黄堇的油质体质量比小于小花黄堇的油质体质量比,讨论了种子特征对蚂蚁散布的影响。     

Odor of achlorophyllous plants’ seeds drives seed‐dispersing ants

Seed dispersal by ants is an important means of migration for plants. Many myrmecochorous plants have specialized appendages in their seeds called elaiosome, which provides nutritional rewards for ants, and enable effective seed dispersal. However, some nonmyrmecochorous seeds without elaiosomes are also dispersed by ant species, suggesting the additional mechanisms other than elaiosomes for seed dispersal by ants. The seeds of the achlorophyllous and myco‐heterotrophic herbaceous plant Monotropastrum humile are very small without elaiosomes; we investigated whether odor of the seeds could mediate seed dispersal by ants. We performed a bioassay using seeds of M. humile and the ant Nylanderia flavipes to demonstrate ant‐mediated seed dispersal. We also analyzed the volatile odors emitted from M. humile seeds and conducted bioassays using dummy seeds coated with seed volatiles. Although elaiosomes were absent from the M. humile seeds, the ants carried the seeds to their nests. They also carried the dummy seeds coated with the seed volatile mixture to the nest and left some dummy seeds inside the nest and discarded the rest of the dummy seeds outside the nest with a bias toward specific locations, which might be conducive to germination. We concluded that, in M. humile seeds, volatile odor mixtures were sufficient to induce seed‐carrying behavior by the ants even without elaiosomes.     

Impact of seed abundance on seed processing and dispersal by the red ant Myrmica rubra

1. Myrmecochory sensu stricto is an ant–plant mutualism in which non‐granivorous ants disperse plant diaspores after feeding on their nutrient‐rich seed appendage, the elaiosome. Phenological traits associated with the diaspore can influence the behaviour of ants and thus their ultimate efficiency as seed dispersers. 2. This study investigated how a contrasting availability of seeds (20 vs. 200 seeds) from the diplochorous Chelidonium majus (Papaveraceae, Linnaeus) plant species influences the behaviour of Myrmica rubra (Formicidae, Linnaeus) ants, from the retrieval of seeds until their dispersal outside the ant nest. 3. Regardless of seed abundance, the ants collected the first diaspores at similar rates. Then, seed retrieval sped up over time for large seed sources until satiation took place with only one‐third of the tested colonies wholly depleting abundant seed sources. 4. No active recruitment by trail‐laying ants was triggered, even to an abundant seed source 5. In both conditions of seed abundance, the majority of the diaspores retrieved inside the nest were discarded with the elaiosome removed and were dispersed at similar distances from the nest. 6. The paper concludes with a discussion of how the quantity of seeds released by a plant with a dual mode of dispersal can potentially influence the behaviour of ant dispersers and hence the dispersal efficiency derived from myrmecochory.     

Ant behaviour and seed morphology: a missing link of myrmecochory

Seed dispersal by ants (myrmecochory) is mediated by the presence of a lipid-rich appendage (elaiosome) on the seed that induces a variety of ants to collect the diaspores. When seeds mature or fall onto the ground, these ant species transport them to their nest. After eating the elaiosome, the seed is discarded in nest galleries or outside, in the midden or farther away, where seeds can potentially germinate. The final location of seeds with their elaiosomes removed was evaluated to assess the importance of possible handles (structures that ants can grasp to carry) in transporting ants during re-dispersal experiments of seeds from nests of six species of ants. The results indicate that seeds remained within the nest because the ants were not able to transport them out of the nest. As a consequence of the elaiosome being removed, small ant species could not take Euphorbia characias seeds out of their nests. Only large ant species could remove E. characias seeds from their nests. Attaching an artificial handle to E. characias seeds allowed small ant species to redistribute the seeds from their nests. On the other hand, Rhamnus alaternus seeds that have a natural handle after the elaiosome removal were removed from the nests by both groups of ant species. If a seed has an element that acts as a handle, it will eventually get taken out of the nest. The ants’ size and their mandible gap can determine the outcome of the interaction (i.e. the pattern of the final seed shadow) and as a consequence, could influence the events that take place after the dispersal process.     

New mutualism for old: indirect disruption and direct facilitation of seed dispersal following Argentine ant invasion

The indirect effects of biological invasions on native communities are poorly understood. Disruption of native ant communities following invasion by the Argentine ant (Linepithema humile) is widely reported to lead indirectly to the near complete collapse of seed dispersal services. In coastal scrub in southeastern Australia, we examined seed dispersal and handling of two native and two invasive alien plant species at Argentine ant-invaded or -uninvaded sites. The Argentine ant virtually eliminates the native keystone disperser Rhytidoponera victoriae, but seed dispersal did not collapse following invasion. Indeed, Argentine ants directly accounted for 92% of all ant-seed interactions and sustained overall seed dispersal rates. Nevertheless, dispersal quantity and quality among seed species differed between Argentine ant-invaded and -uninvaded sites. Argentine ants removed significantly fewer native Acacia retinodes seeds, but significantly more small seeds of invasive Polygala myrtifolia than did native ants at uninvaded sites. They also handled significantly more large seeds of A. sophorae, but rarely moved them >5 cm, instead recruiting en masse, consuming elaiosomes piecemeal and burying seeds in situ. In contrast, Argentine ants transported and interred P. myrtifolia seeds in their shallow nests. Experiments with artificial diaspores that varied in diaspore and elaiosome masses, but kept seed morphology and elaiosome quality constant, showed that removal by L. humile depended on the interaction of seed size and percentage elaiosome reward. Small diaspores were frequently taken, independent of high or low elaiosome reward, but large artificial diaspores with high reward instead elicited mass recruitment by Argentine ants and were rarely moved. Thus, Argentine ants appear to favour some diaspore types and reject others based largely on diaspore size and percentage reward. Such variability in response indirectly reduces native seed dispersal and can directly facilitate the spread of an invasive alien shrub.     

Hydrated mucilage reduces post-dispersal seed removal of a sand desert shrub by ants in a semiarid ecosystem

Post-dispersal seed removal by animals can lead to extensive seed loss and thus is an important factor in structuring plant communities. However, we know much less about post-dispersal seed predation than about other forms of herbivory. Mucilage plays many ecological roles in adaptation of plants to diverse environments; nevertheless, until now the role of mucilage in ant-mediated seed movement remains largely hypothetical. We studied the role of mucilage in seed removal of Artemisia sphaerocephala by ants in Mu Us Sandland in Inner Mongolia, China. Messor aciculatus was the most active seed predator of Artemisia sphaerocephala. Time to first ant collecting (T _1st) of wet intact seeds was longest and significantly different from that for dry intact seeds, wet demucilaged seeds, and dry demucilaged seeds; number of seeds removed to ant nests was lowest for wet intact seeds. After they were collected by ants, 5 % of wet intact seeds were dropped during transport. Our results indicate that seed mucilage of Artemisia sphaerocephala may play a significant role in post-dispersal seed removal by (1) making seeds less attractive to ants, thus resulting in a delay of collection time; (2) forming a strong bond to soil particles, making it difficult for ants to remove seeds; and (3) making seeds more likely to be dropped during transport, thereby allowing them to escape from predation even after collection by ants. This study demonstrates the importance of mucilage in reducing seed removal by ants and thus in anchoring seeds of desert plants in the vicinity of mother plants.     

Differential dispersal of Chamaesyce maculata seeds by two ant species in Japan

Seed dispersal by ants (myrmecochory) is a widely distributed plant–animal interaction in many ecosystems, and it has been regarded as a generalized (multiple species) interaction in which specialization on specific ant partners is uncommon. In this paper, we demonstrate species-specific seed dispersal of spotted spurge (Chamaesyce maculata) by ants in Japan. C. maculata produces seeds from summer to autumn in Japan. The seeds produced in autumn are carried by two ant species, Tetramorium tsushimae and Pheidole noda. We performed laboratory experiments to investigate the fate of C. maculata seeds in the nests of T. tsushimae and P. noda. P. noda consumed the seeds in the nest and rarely carried seeds out of the nest, while T. tsushimae consumed only the seed coat, and subsequently carried the seeds out of the nest. Removal of the seed coat by T. tsushimae may increase seed survival by reducing their susceptibility to infection by fungi. We also observed ant responses to filter paper soaked with an aqueous extract of the seed coat. P. noda did not react to the filter paper, but T. tsushimae carried the filter paper into their nest. Analysis by high-pressure liquid chromatography revealed that the aqueous extract contained at least four sugars and one unknown substance. Myrmecochory has been regarded as a generalized interaction with specialization for specific ant partners uncommon. However, our study suggests there is a species-specific interaction in seed dispersal by ants in autumn-flowering individuals of C. maculata in Japan.     

Seed removal decrease by invasive Argentine ants in a high Nature Value farmland

Seed dispersal by ants is an important ecological process that maintains the structure anddiversity of natural communities, however, it is vulnerable to biological invasions. Argentine ants are one of the worst invasive ant species and cause severe changes in ecosystem processes and native ant biodiversity declines in invaded sites. Here, we studied seed removal by ants combining observations and a cafeteria experiment with seeds of four myrmecochorous plant species (Centaurea sphaerocephala, Rosmarinus officinalis, Silybum marianum, and Ulex australis) in two sites (invaded and uninvaded) located in the Mediterranean Montado ecosystem and classified as High Nature Value farmland (HNV). Significant differences in daily seed removal rates were found between the two study sites. In uninvaded sites, several native ant species were attracted to the seeds, resulting in all seeds being removed rapidly. The majority of seed removal events were carried out by two key seed disperses Pheidole pallidula (71%) and Aphaenogaster iberica (26%) with a clear preference for diaspored with larger and heavier elaiosome (i.e., C. sphaerocephala, S. marianum). By contrast, while the Argentine ant showed some interest (68% of seeds were interacted with), no seed removal events were observed. The extirpation of the local ant fauna by the Argentine ant and its inability to ensure seed dispersal services may lead to the interference and eventually to the collapse of seed dispersal of the four studied myrmecochorous plants in the invaded site in the future. We argue that these discrete but severe consequences of an invasive species on a key ecological process may strongly affect the functioning of the Montado ecosystem.     

Edge effects decrease ant-derived benefits to seedlings in a neotropical savanna

Edge effects may lead to changes in mutualistic plant–animal interactions, such as seed dispersal, that are critical to plant regeneration. However, research into edge effects is neglected in the Brazilian cerrado, the largest neotropical savanna. We evaluated the consequences of edge effects in the cerrado for the regeneration of Erythroxylum pelleterianum (Erythroxylaceae), a shrub that benefits from seed dispersal by ants. We compared air temperature, relative humidity, and vapor pressure deficit, as well as the frequency and outcome of ant–diaspore interactions between cerrado edges and interiors. The inner portion of cerrado was likely to be moister than its borders, but seed production and germination did not differ between edge and interior of cerrado. Ants removed more seeds near fragment edges than at the interior. However, Myrmicinae ants dominated ant–fruit interactions at edges. These ants are likely to provide few benefits to the seeds. Seedlings of E. pelleterianum growing close to Ponerinae ant nests showed higher survival than seedlings growing away from nests in the interior of cerrado, but such effect disappeared near edges. Widespread seedling mortality due to a higher evaporative demand at edges may partially account for this effect. Furthermore, Ponerinae’s nests also showed a lower residence time near edges, decreasing possible benefits derived from ant colony activity such as nutrient enrichment and protection against insect herbivores. Edge effects could change the structure and dynamics of vegetation in cerrado fragments, due in part to the collapse of the mutualistic interactions demonstrated here.     

Combining distances of ballistic and myrmecochorous seed dispersal in Adriana quadripartita (Euphorbiaceae)

The separate contributions of different vectors to net seed dispersal curves of diplochorous systems have rarely been characterised. In Australia, myrmecochory is a common seed dispersal syndrome and in the majority of such systems, seeds are initially dispersed ballistically. We measured ballistic and myrmecochorous seed dispersal distances in relation to canopies of Adriana quadripartita (Euphorbiaceae) and used a simulation model to estimate the net dispersal curve. We also compared seed removal rates and ant abundances under, and outside, plant canopies to examine how foraging patterns by ants may affect net dispersal.Overall ant abundance did not show a significant numerical response to seedfall; however, the abundance of the main seed dispersing ant, Rhytidoponera ‘metallica’ did. Despite this, seed removal rates did not differ significantly between canopy and open locations. Rhytidoponera ‘metallica’ account for 93% of observed seed dispersal events. On average, the ants dispersed seeds 1.54 m and in doing so, moved seed a mean radial distance of 0.76 m away from canopy edges. This contribution to net dispersal distance by ants is considerable since ballistic dispersal moved seeds a median distance of 7.5 cm. Our simulation model indicated that the combination of ballistic and ant seed dispersal is expected to result in seeds being transported a median net radial dispersal distance of 1.05 m from the canopy edge.Thus in this system, an important function of diplochory may simply be to move a higher proportion of seeds from under the canopy of parent plants than is possible by ballistic dispersal alone. This ‘dispersal-for-distance’ may result in reduced parent–offspring competition or may increase the probability that seeds reach rare safe sites for germination and recruitment.     

Seed dispersal by ants in the semi-arid Caatinga of North-East Brazil

BACKGROUND AND AIMS: Myrmecochory is a conspicuous feature of several sclerophyll ecosystems around the world but it has received little attention in the semi-arid areas of South America. This study addresses the importance of seed dispersal by ants in a 2500-km(2) area of the Caatinga ecosystem (north-east Brazil) and investigates ant-derived benefits to the plant through myrmecochory. METHODS: Seed manipulation and dispersal by ants was investigated during a 3-year period in the Xingó region. Both plant and ant assemblages involved in seed dispersal were described and ant behaviour was characterized. True myrmecochorous seeds of seven Euphorbiaceae species (i.e. elaiosome-bearing seeds) were used in experiments designed to: (1) quantify the rates of seed cleaning/removal and the influence of both seed size and elaiosome presence on seed removal; (2) identify the fate of seeds dispersed by ants; and (3) document the benefits of seed dispersal by ants in terms of seed germination and seedling growth. KEY RESULTS: Seed dispersal by ants involved one-quarter of the woody flora inhabiting the Xingó region, but true myrmecochory was restricted to 12.8 % of the woody plant species. Myrmecochorous seeds manipulated by ants faced high levels of seed removal (38-84 %) and 83 % of removed seeds were discarded on ant nests. Moreover, seed removal positively correlated with the presence of elaiosome, and elaiosome removal increased germination success by at least 30 %. Finally, some Euphorbiaceae species presented both increased germination and seedling growth on ant-nest soils. CONCLUSIONS: Myrmecochory is a relevant seed dispersal mode in the Caatinga ecosystem, and is particularly frequent among Euphorbiaceae trees and shrubs. The fact that seeds reach micro-sites suitable for establishment (ant nests) supports the directed dispersal hypothesis as a possible force favouring myrmecochory in this ecosystem. Ecosystems with a high frequency of myrmecochorous plants appear not to be restricted to regions of nutrient-impoverished soil or to fire-prone regions.     

A stochastic model of seed dispersal pattern to assess seed predation by ants in annual dry grasslands

A combined field experiment and modelling approach has been used to provide evidence that ants may be responsible for an observed lower patchiness and higher plant diversity in the neighbourhood of ant nests, within Mediterranean dry grasslands belonging to the phytosociological class Tuberarietea guttatae. The hypothesis was that seeds occurring in clumps may have a higher probability to be harvested than seeds having a scattered distribution. In order to test this hypothesis, four analysis steps were performed. First, pattern of seed production and dispersal of four species was recorded; two of them were more abundant next to ant nests (Tuberaria guttata, Euphorbia exigua), whereas the other two were more abundant away from ant nests (Bromus scoparius and Plantago bellardi). Second, a stochastic model was developed to simulate the observed dispersal patterns of each studied species. Third, 10 seed spatial arrangements in accordance to the distribution patterns created by the model were offered to ants and the location of predated seeds was recorded. Finally, the observed pattern of seed predation was matched to models performed by different distributions of probability. Results showed that the probability of being predated decreased as distance among seeds increased. This preference of ants for high concentration of food items holds down the dominant species sufficiently to allow the subordinates to survive, thus increasing diversity near nests. The observed higher frequency of small-seeded, small-sized, or creeping therophytes close to the ant nests can be therefore seen as an example of indirect myrmecophily.     

Are red imported fire ants facilitators of native seed dispersal?

Invasive ants threaten native communities, in part, through their potential to disrupt mutualisms, yet invasive species may also facilitate native species. The red imported fire ant (Solenopsis invicta) is one of the most conspicuous invasive ants in North America and its high densities, combined with its potential to displace native ants, have led to concerns that it may disrupt ant-plant seed dispersal mutualisms. We examined the potential of fire ants to disperse seeds in the longleaf pine ecosystem by comparing the removal of elaiosome-bearing seeds by fire ants versus native ants. A total of 14 ant species were observed removing seeds, with fire ants responsible for more than half of all removals. While fire ants were the dominant seed remover in this system, they did not remove significantly more seeds than would be expected based on their population density (46% of ground-dwelling ants). Moreover, red imported fire ants were similar to native ants with respect to distance of seed movement and frequency of moving seeds back to the nest. Areas of higher fire ant densities were found to have greater rates of seed removal by ants without a subsequent drop in seed dispersal by native ants, suggesting that fire ant-invaded areas may experience overall higher levels of seed dispersal. Thus, fire ants may actually facilitate dispersal of elaiosome-bearing plant species in the longleaf pine ecosystem.     

Myrmecochory and short-term seed fate in Rhamnus alaternus: Ant species and seed characteristics

Benefits conferred on plants in ant-mediated seed dispersal mutualisms (myrmecochory) depend on the fate of transported seeds. We studied the effects of elaiosome presence, seed size and seed treatment (with and without passage through a bird's digestive tract) on short-term seed fate in Rhamnus alaternus. In our study, we define short-term seed, or initial, seed fate, as the location where ants release the seeds after ant contact with it. The elaiosomes had the most influence on short-term fate, i.e. whether or not seeds were transported to the nest. The workers usually transported big seeds more often than small ones, but small ants did not transport large seeds. Effect of seed size on transport depended on the ant species and on the treatment of the seed (manual extraction simulating a direct fall from the parent plant vs. bird deposition corresponding to preliminary primary dispersal). Probability of removal of elaiosome-bearing seeds to the nest by Aphaenogaster senilis increased with increasing seed weight.     

Signals Can Trump Rewards in Attracting Seed-Dispersing Ants

Both rewards and signals are important in mutualisms. In myrmecochory, or seed dispersal by ants, the benefits to plants are relatively well studied, but less is known about why ants pick up and move seeds. We examined seed dispersal by the ant Aphaenogaster rudis of four co-occurring species of plants, and tested whether morphology, chemical signaling, or the nutritional quality of fatty seed appendages called elaiosomes influenced dispersal rates. In removal trials, ants quickly collected diaspores (seeds plus elaiosomes) of Asarum canadense, Trillium grandiflorum, and Sanguinaria canadensis, but largely neglected those of T. erectum. This discrepancy was not explained by differences in the bulk cost-benefit ratio, as assessed by the ratio of seed to elaiosome mass. We also provisioned colonies with diaspores from one of these four plant species or no diaspores as a control. Colonies performed best when fed S. canadensis diaspores, worst when fed T. grandiflorum, and intermediately when fed A. canadense, T. erectum, or no diaspores. Thus, the nutritional rewards in elaiosomes affected colony performance, but did not completely predict seed removal. Instead, high levels of oleic acid in T. grandiflorum elaiosomes may explain why ants disperse these diaspores even though they reduce ant colony performance. We show for the first time that different elaiosome-bearing plants provide rewards of different quality to ant colonies, but also that ants appear unable to accurately assess reward quality when encountering seeds. Instead, we suggest that signals can trump rewards as attractants of ants to seeds.     

DUAL FUNCTION OF THE ELAIOSOME OF CORYDALIS AUREA (FUMARIACEAE): ATTRACTION OF DISPERSAL AGENTS AND REPULSION OF PEROMYSCUS MANICULATUS,A SEED PREDATOR

It has been suggested that one of the selective advantages of ant dispersal is the burial of seeds in ant nests where predators such as small rodents cannot find them. The elaiosomes of Corydalis aurea (Fumariaceae) are extremely attractive to ants, which assiduously gather the seeds and take them to nests. However, seed production commonly exceeds the gathering capacity of ants so that seeds accumulate beneath the parent plants. In spite of this, no signs of rodent predation are evident. Experiments with a major seed predator, the deer mouse Peromyscus maniculatus, show that when given the choice of seeds with and without elaiosomes, the mice consume significantly more seeds without elaiosomes. This remains true whether or not the intact seeds bear fresh, moist elaiosomes or dry, withered ones. Our experiments strongly suggest that the elaiosome has a dual function, the attraction of the ant seed dispersers and the repulsion of seed predators.     

Composition of harvested seeds and seed selection by the invasive tropical fire ant, <Emphasis Type="Italic">Solenopsis geminata</Emphasis> (Hymenoptera: Formicidae) in Taiwan

Solenopsis geminata (F.) is an invasive ant that is widely distributed in weedy areas and agricultural fields in Taiwan. Previous studies have found that S. geminata harvests the seeds of numerous plants. In the present study, we further investigated the composition of harvested seeds in ant nests and seed selection by workers. The seed caches in S. geminata nests sampled in eight areas in Taiwan suggest that the seeds harvested by workers were diverse and belonged to 52 plant species in 17 plant families. Twenty-three species (44%) belong to the family Gramineae, and most of the seeds weighed from 0.02 to 2.29 mg, which might suggest that these are the main seeds harvested by S. geminata from their habitat. Ten common species with similar seed weights were used to compare the seed preferences of workers from two experimental sites. The results suggest that the seed preference was different between the two experimental sites. The seeds of Casuarina equisetifolia showed the most obvious difference in seed removal speed, which might suggest that S. geminata potentially prefers the encountered seed species in the habitat. The various plant species in the ant nests and seed preference suggest that fire ants easily accept newly encountered plant species. As more than half of the plant species (52%) and the total seed number (63%) belonged to exotic species, the role of S. geminata might be negative because it tends to harvest exotic seeds and has a high opportunity to improve the establishment of exotic seedlings.