DYNAMICS OF SOIL-STORED SEED BANKS OF 6 SHRUBS IN FIRE-PRONE DUNE FYNBOS

(1) Soil-stored seed banks of six non-sprouting dominant shrubs with small seeds (0.4-4.6 mg) were determined in three mature (13-year-old) dune fynbos communities each autumn for 3 successive years. A comparison of methods of estimating seed banks by counting seeds vs. counting germinants showed neither method to be superior; a knowledge of other life-history characteristics, such as seed input and germination cues, was essential in evaluating results. (2) Instead of the large seed banks typical of shrubs from fire-prone vegetation, dune fynbos shrubs had relatively small seed banks: the highest mean seed density recorded was 825 +/- 1187 seed m-2 (mean +/- 1S.D., n = 50). Shrub and community seed banks showed no significant increments over 3 years. After a fire, seed banks were depleted by more than half. (3) Removal of seeds by ants and rodents from the soil surface was relatively rapid. Almost all seed had been removed from open depots by the sixth day of cafeteria experiments. Removal rates by ants of elaiosome-bearing (myrmecochorous) seeds was not significantly different from non-myrmecochorous species; neither were the intact seeds of a myrmecochorous species removed faster than the seeds from which the elaiosome had been removed. The results indicate that seeds are removed by ants as much for granivory as for myrmecochory. (4) Seed burials showed that seed bank losses due to decay and germination varied widely amongst species and with seed age, with losses ranging from 12% to 93%. (5) There were large differences in seed-bank sizes between the six species which could be attributed to their life-history attributes such as seed production, dispersal mode and germination requirements. The myrmecochorous species with relatively large and poorly dispersed seeds, small to moderate seed production and seed-bank sizes, and germination and seedling establishment cued to post-fire conditions, were most vulnerable to local extinction.