This study has examined the rate at which the seeds of certain species disappear from woodland soils, and whether the composition of the seed banks beneath stands of known age can be used to reveal management history. Associated with this is an investigation of the spatial distribution of the stored propagules. A question arises as to whether species with long lived seeds can survive as buried seed long enough to take advantage of the opening of the canopy and soil disturbance resulting from the falling of old trees. Given the great life span of most trees, this is a strategy open only to species with exceptionally long lived seed banks. One such species is Hypericum pulchrum, which, as the seed bank surveys carried out in this study show, can exist as viable seeds beneath woodlands of great age in the Tavistock Woodland Estate in Devon. Is this germination and reproduction on tree falls sufficient to maintain Hypericum in the seed bank indefinitely ? To answer this question, a computer model was developed to examine the effect of various parameters on the survival time of a Hypericum seed bank beneath a simulated woodland. The model allowed prediction of both the mean density of buried seeds in a unit area, and the development of pattern within the seed bank. Analysis of the simulation results established that the dispersal of Hypericum was insufficient to maintain a seed bank by a "chain reaction" of windthrown reinforcement. The implications of this are discussed with respect to pioneer species within both temperate and tropical forests.