SOIL SEED BANK DYNAMICS IN RELATION TO TOPOGRAPHIC POSITION OF A MIXED-DECIDUOUS FOREST IN SOUTHERN NEW-ENGLAND, USA

We examined the floristic significance of soil seed banks in relation to valley, midslope and ridge sites in a 70-90-year old forest in nort heastern Connecticut. A-horizon mineral soils were collected to 5 cm d epth in the early spring from forest understory sites across the topog raphy. Samples from each of the sites were exposed to full sun within a greenhouse. Records of germination were made at regular intervals ov er a 60-day period. These showed significant differences among sites i n number of species and total number of germinants. Greatest numbers o f species and germinants were recorded from valley sites and these pro gressively declined from midslope to ridgetop. Twenty-five different s pecies were identified. Species were grouped into growth habits - gram inoids, herbs, shrubs, trees, and vines. Over 61% of all germinants ac ross all sites were graminoids. Seventy-four percent of all germinants in the valley sites were graminoids, with over 93% of them represente d by two Sedge species, Carer glaucodea and C. lupulina. On the ridget op sites graminoids were more evenly distributed among six different s pecies. The percentage in each growth habit changed rank across topogr aphic position with germinants of graminoids and trees most abundant o n valley sites; herbs, on midslopes, and shrubs, on ridgetops. All ger minants, except for those of the trees, Carer spp., and two herb speci es were weedy species that were not characteristic of the existing veg etation. Germinants of the trees, Carer spp., and herbs that were char acteristic of the existing vegetation were mostly confined to soils fr om the valley sites. Two weedy herbs, Plantago major and Verbascum tha psus, are exotic introductions that originally came from Europe. Only one vine, Vitis aestivalis, from a midslope site germinated. Species d iversity is higher on midslope sites than valleys and ridgetops. The s ignificance of these findings in relation to site productivity and dis turbance history is discussed. (C) 1998 Elsevier Science B.V.