No abstract

Seed re‐dispersal following initial harvesting by ants may have important implications for the distribution and fate of myrmecochorous seeds. However, the probability of seed re‐dispersed by ants and the effect it may have on subsequent survival appear variable, the functional role of diaspore, disperser and seed predator to the fate of discarded seeds remain unclear. To clarify the ecology, we compared the consequences of seed re‐dispersal by a keystone seed‐dispersing ant (Myrmica ruginodis Nylander) for four sympatric myrmecochorous plants common to the temperate deciduous forests in Qinling Mountains, central China. Plants varied in the probability of re‐dispersal and in elaiosome condition. Ants preferred seeds with residual elaiosomes, while rodents only consumed the two larger‐seeded species, regardless of the elaiosome presence. The scattered distribution of discarded seeds increased the probability of ant re‐harvesting and, to some extent, reduced rodent predation. Thus, difference in the probability of seed re‐dispersal and its subsequent effect on seed fate in relation to ants and rodents was attributed primarily to the elaiosome condition, seed size and seed spatial pattern. The results imply that seed re‐dispersal could affect the fitness of plants and ultimately influence the plant abundance and distribution pattern. This highlights the necessity to incorporate re‐dispersal into myrmecochory to advance our understanding of the benefits of myrmecochory to plants.

Seed dispersal in myrmecochorous plants, in which a sequence of two or more steps is often involved, has profound effects on plant benefits. First, seeds mature or fall onto to the ground (primary dispersal), then ants transport them into nests (ant dispersal), and sometimes ants subsequently discard seeds out of nests (re‐dispersal). By neglecting one or the other of these phases, we might misjudge the benefits of myrmecochory to plants. Here, we investigated the difference between two common Chinese myrmecochorous plants, Corydalis incisa and C. wilfordii, in the importance of each dispersal step by examining 1) primary dispersal distance, 2) seed removal rates and dispersal distance by ants, and 3) frequency of seeds discarded out of nests and the re‐dispersal distance. We found that the mean primary dispersal distance of C. incisa was about eight times longer than that of C. wilfordii. The presence of an elaiosome increase the attractiveness of seeds to ants, and both the removal rate and dispersal distance were greater in C. wilfordii than in C. incisa. Two ant species, Pheidole noda and Pristomyrmex pungens are the major dispersers. Pheidole noda, the larger‐bodied ant species, discarded most seeds of both species out of nests, whereas Pr. pungens discarded only C. wilfordii seeds. The mean re‐dispersal distances of the two plant species were similar, but that of C. incisa and C. wilfordii were about threefold and fivefold shorter than the distances in the step of ant dispersal, respectively. In the field, no predators were found to interact with the discarded seeds in 72 h. Our results indicate that the importance of each dispersal phase differ between the two Corydalis species, and this difference is largely attributable to the differential role of ant identity and seed characteristics. It is concluded that re‐dispersal also needs to be considered in studies of myrmecochory.