Abstract
The composition of a forest seed bank was estimated using two methods: (i) seed extraction, i.e., the physical separation of the seeds from the soil via flotation in a salt solution, and (ii) seedling emergence, i.e., the germination of seedlings from soil samples incubated under greenhouse conditions for 5 months. The extraction method predicted a density of 12 500 seeds∙m−2, while the emergence method detected 3800 émergents∙m−2. There was considerable disparity in species composition derived from the two methods. The extraction method identified 102 different taxa, with 22 species making up 99% of the seeds and 5.6 + 0.2 species per sample. In contrast, the emergence technique identified fewer species (60) but had more species per sample (7.6 + 0.2). Eleven species made up 99% of the emergents. Verbascum thapsus represented 34% of the seedlings in the emergence study but only 1 % of the extracted seeds. Members of the Polygonaceae represented 19% of the extracted seeds but less than 1 % of the seedling emergents. No tree or shrub species were found with the emergence method, although they represented 8% of the extracted seeds. There was a poor correlation between the estimates of species number, seed density, and diversity obtained from the two methods. The seed extraction method had considerably higher variability for these parameters. It is apparent from this study that the seedling emergence and seed extraction methodologies do not produce similar estimates of the seed bank composition. The differences are such that comparisons should not be drawn between studies using the different methods. Careful considerations should be given to both the objectives of the seed bank study and the relevant literature prior to the selection of an appropriate method. Key words: seed bank, method, composition, diversity, density, sample number.
Publisher
Canadian Science Publishing