Seed re-dispersal of four myrmecochorous plants by a keystone ant in central China

Ant Guild Identity Determines Seed Fate at the Post-Removal Seed Dispersal Stages of a Desert Perennial

Simple Summary

Ants play a dual role in their interaction with plant seeds. Many ant species, mainly harvester ants, consume plant seeds (granivory), whereas other ants, mainly scavengers, provide a beneficial seed dispersal service. Granivory by ants is frequently documented in deserts, but beneficial seed dispersal is rarely studied in such ecosystems. We followed the handling of seeds of Sternbergia clusiana, an ant-dispersed plant, by two guilds of desert ants. We focused on the treatment of seeds within the nest and on the redispersal of seeds after ejection from the nest. Scavenging ants rarely consumed the seed itself, they deposited most seeds away from the nest in apparently suitable microhabitats. In addition, we found that most of the seeds that were relocated by scavenging ants arrived at sites under the canopy of shrubs. Such sites might be beneficial for the establishment and success of plants in the arid environment. Indeed, we found that the subject plant was more likely to be found under shrubs than elsewhere, which suggests that it benefits from being placed there. Such documentation of seed dispersal activity by scavenging ants in arid ecosystems suggests they may be efficient seed dispersers in deserts, as they are in other ecosystems.

Abstract

Ants play a dual role in their interaction with plant seeds. In deserts, the consumption of seeds by granivorous ants is common, whereas mutualistic seed dispersal, often associated with scavenging ants, is rarely documented. We evaluated the contribution of both ant guilds to efficient seed dispersal of an ant-dispersed plant, Sternbergia clusiana, in a desert ecosystem. We presented seed to colonies of three species of desert ants from the Cataglyphis (scavengers) and Messor (granivorous) genera. We recorded seed consumption, ejection from the nest, and seed transportation to potentially beneficial microhabitats. We evaluated microhabitat quality by testing the association between habitat types and the plant at various life stages. As expected, granivores mainly consumed the seeds, whereas scavengers consumed the elaiosome (seed appendage serving as a reward), but left the seeds intact. Moreover, scavenging ants relocated the seeds much further than granivores, mainly to shrub patches. The disproportional distribution of the plant under shrubs at several life stages suggests that this microhabitat is beneficial for the plant. Overall, while granivores seem to mainly harm seed dispersal, we provide the first evidence for the beneficial contribution of scavenging ants in deserts, showing they exhibit the same suite of characteristics that render them efficient seed dispersers in other ecosystems.

Interactions between seed-dispersing ant species affect plant community composition in field mesocosms

In generalized mutualisms, species vary in the quality of services they provide to their partners directly via traits that affect partner fitness and indirectly via traits that influence interactions among mutualist species that play similar functional roles. Myrmecochory, or seed dispersal by ants, is a generalized mutualism with ant species varying in the quality of dispersal services they provide to their plant partners. Variation in ant species identity can directly impact seed dispersal patterns and plant community composition; however, we know less about how interactions among seed-dispersing ant species indirectly influence plant partners. The invasive ant Myrmica rubra, is a high-quality seed-disperser in its native range that interacts with myrmecochores (ant-dispersed plants) and the high-quality seed disperser Aphaenogaster sp. in its invaded range. We use this system to examine how interactions between two functionally similar mutualist ant species influence the recruitment and community composition of ant-dispersed plants. We performed a field mesocosm experiment and a laboratory behavioural experiment to compare discovery and dominance behaviours between ant species, and seed dispersal and seedling recruitment of four myrmecochore species among intraspecific interaction treatments of each ant species and an interspecific interaction treatment. We found that M. rubra was better at discovering and dispersing seeds, but Aphaenogaster sp. was dominantly aggressive over M. rubra. Interspecific interactions dampened seed dispersal relative to dispersal by the better disperser. Despite this dampening, we found no effect of interspecific interactions on seedling recruitment. However, community composition of seedlings in the interspecific interaction treatment was more similar to composition in the aggressively dominant ant (Aphaenogaster sp.) treatment than in the better discoverer ant M. rubra treatment. We show that interspecific interactions between mutualist species in the same functional guild affect the outcome of mutualistic interactions with partner species. Despite the native ant dispersing fewer seeds, its dominance over the subordinate (invasive) ant has the potential to allow for some level of biotic resistance against the effects of M. rubra on plant communities when these species coexist.

Invasive ants disperse seeds farther than native ants, affecting the spatial pattern of seedling recruitment and survival

Mutualists can vary in the quantity and quality of service which they provide to their partners. Variation in seed disperser quality depends on seed-processing traits, dispersal distance, and deposition location, all of which ultimately affect plant fitness. Here, we compared these aspects of seed dispersal quality between a native and an invasive ant species, and examined how they affect competition and plant performance. Using experimental mesocosm communities, we examined how these two ant species affect the spatial pattern of recruitment and establishment for four myrmecochorous plant species, including one invasive species. We measured the locations of dispersed seedlings relative to ant nests, adult plants, and other dispersed seedlings, as well as measured the effects of location on plant performance. The invasive ant, Myrmica rubra, secondarily dispersed seeds farther from its nests, creating a less clumped pattern of seedling recruitment compared to the native ant, Aphaenogaster rudis. Plant species responded differently to dispersal. Invasive seedlings recruited farther from adult plants than native seedlings, and had higher survival the farther they were from conspecifics. In contrast, native plants had higher survival and grew taller when dispersed farther from invasive plants. We show that seed-dispersing ant partners differ in mutualist quality creating differences in dispersal distance and deposition location that affects a plant's competitive environment. Our results reveal the potential for long-term consequences on plant community structure with changing ant partner identity. We emphasize the need to examine dispersal quality in addition to quantity to uncover the importance of partner identity in structuring communities.

Investment in reward by ant-dispersed plants consistently selects for better partners along a geographic gradient

Seed dispersal by ants (myrmecochory) is an asymmetric, presumably mutualistic interaction, where a few ant species benefit many plants. Myrmecochorous plants express specialized adaptations, most notably a large elaiosome, which promote interactions with efficient seed dispersers while decreasing interactions with poor dispersers, resulting in de facto partner choice. However, because variation in plants' investment in reward and ant response to them may vary spatially and temporally, it is unclear whether such specialization is consistent along geographic gradients; especially towards myrmecochory's range margin. To answer this question on context-dependent partner choice, we first estimated variation in reward investment by co-occurring myrmecochores along a steep environmental gradient in a Mediterranean region. Second, we tested whether variation in plant investment in reward was positively and consistently correlated with the quality of dispersal plant received along the same gradient. Using in situ cafeteria experiments, we simultaneously presented diaspores of locally co-occurring myrmecochorous species to ants of two guilds representing high- and low-quality dispersers. We then recorded ant-seed behaviour, seed preference and seed removal rates for each ant guild. We found both overall and within-site high variation among plant species in the total and relative investment in elaiosomes. Both ant guilds removed substantial proportion of the seeds. However, scavenging ants (high-quality dispersers) clearly preferred diaspores with larger elaiosomes, whereas granivorous ants (low-quality dispersers) exhibited no preference. Furthermore, both the variation in plant traits and the corresponding response of different ant guilds were consistent along the studied geographic gradient. This consistency holds even when granivores, which removed seeds in a non-selective fashion and provided apparently low-quality seed dispersal services, were, at least numerically, the dominant ant guild. This dominance and the consistency of the partner choice shed light on the functionality of elaiosomes at the margins of myrmecochory's distribution.