Diplochory in Viola: A Possible Relation between Seed Dispersal and Soil Seed Bank

Abstract

Patterns of Viola seeds on the ground after ballistic dispersal, patterns of ant search intensity for seeds, and ant attraction to piles of different size of Viola seeds were assessed. Density of seeds following ballistic dispersal decreased with increasing distance from the parent plant. Search intensity varied between adjacent small areas, the environmental grain size being larger than 3 cm in three out of four localities. Each seed was more likely to be taken by ants if it was located close to another seed, or in a relatively large pile (2-9 seeds per pile). The probability of a Viola seed being taken by an ant in the field was highly variable among species and to some degree determined by the distance of the ballistic dispersal. Diplochory was interpreted as a mechanism which may result in a double seed bank, with some seeds being carried to ant nests whereas other seeds are left on the soil surface and may become incorporated into the soil seed bank outside ant nests. INTRODUCTION Many myrmecochorous plant species with diplochory (ballistic dispersal prior to ant dispersal) have been identified since the beginning of this century (see Sernander, 1906; Bresinsky, 1963; Beattie and Lyons, 1975; Handel, 1978; Handel et al., 1981; Westoby et al., 1982). This type of dispersal is energetically expensive. To initially scatter seeds widely by ballistic dispersal requires long pedicels and strong capsules. To also produce energy-rich elaiosomes, which later attract ants to the seeds so that the ants collect the seeds in the ant nests, requires an additional energy investment. If maximizing the dispersal distances of the seeds were the main reason for diplochory, the two methods of dispersal should reach approximately the same distances for most efficient dispersal (Westoby and Rice, 1981). Many seeds, however, will be located closer to the parent plant by diplochory than by ballistic dispersal alone (see Beattie and Culver, 1981). Ants gradually deplete seed pools on the soil surface. Seeds lying closest to the ant nests are taken preferentially to seeds lying further away in areas dominated by seed harvester ants (DeVita, 1979). The search intensity for seeds also varies between sites 10 m apart (Anderson, 1982), but there are no data on the probability for individual seeds to be taken by ants within smaller areas. Many Viola species in the temperate zones have ballistic dispersal followed by ant dispersal. Beattie and Lyons (1975) have suggested that this could be a way to avoid predation on the capsules by slugs, while Culver and Beattie (1977) interpreted the ant dispersal as a way to avoid predation by rodents and butterfly larvae. Further, Culver and Beattie (1980) indicated that Viola seed germination was increased for seeds that were planted in ant nests compared to seeds in the surrounding soil. They suggested that the elaiosomes might be a vehicle for the seeds to reach sites suitable for germination. As it seemed intuitively an uneconomical use of energy for a plant to invest in both ballistic dispersal and ant dispersal, the following questions were addressed to investigate the influence of ballistic dispersal on the subsequent ant dispersal: (1) What pattern is formed on the ground by the seeds after ballistic dispersal of seeds of three different Viola species? (2) Is the search intensity of ants evenly intensive, or are some areas searched more intensively than nearby areas? (3) Is the probability for individual 1 Present address: Botanical Institute, Veterinary & Agricultural University, 23 Rolighedsvej, DK-1958 Copenhagen V, Denmark.