Influence of panicle fragment size on ant seed-harvesting and seedling recruitment in Phalaris aquatica

Abstract

Successful seedling recruitment events in sparse stands of the grass Phalaris aquatica (phalaris) are rare. A major contributing factor to this situation is the seed-harvesting activity of ants. In recently developed cultivars of phalaris, the seed is retained within the glumes of the floret and is shed in a seed-retaining panicle fragment. We tested the hypothesis that seed contained in panicle fragments was less likely to be harvested by ants than naked seed and that this would be reflected in increased seedling recruitment. Observations of ant seed-harvesting activity and records of seedling recruitment following dispersal of seed and panicle fragments onto field plots were made at Canberra, Australian Capital Territory, and Tamworth, New South Wales, over 2 years. In the second year, the effects of 2 pasture renovation treatments (sheep-grazing and shallow-discing) on seedling recruitment from seed and panicle fragments were examined. Averaged over the 2 localities, 4% of large-sized (>10 florets), 13% of medium-sized (5-10 florets) and 27% of the small-sized (2 or 3 florets) panicle fragments were harvested in a 24 h period by ants (mainly Pheidole spp.), compared with 83% of the naked phalaris seed. The effects of panicle fragment size on recruitment were variable depending on site in the first year and the interaction of site and renovation treatments in the second year, but contrary to expectation, recruitment often was not higher in panicle fragment treatments than in seeded treatments. Seedling recruitment at Canberra (a hardsetting sandy loam) was lower (22 plants/m² from naked seed and 15 plants/m² from seed in panicle fragments) because of the competition from existing vegetation and germination of subterranean clover from the seedbank, than at Tamworth (a self-mulching red chromosol; 179 plants/m² from naked seed and 118 plants/m² from seed in panicle fragments), where seedling recruitment was also significantly increased by the sheep-grazing renovation treatment. At Tamworth, a larger seedbank was recovered from plots of oversown panicle fragments compared with plots oversown with naked phalaris seed. This suggested that protection of seeds in panicle fragments over the summer months on the Tablelands of New South Wales would increase the time over which phalaris seeds are available for recruitment. Further work is needed to test seedling recruitment in panicle shattering v. non-shattering phalaris populations and to define the conditions under which the seedbank contained in the panicle fragments can be exploited to improve stand persistence.