The most often used time-line for distinguishing a transient seed bank from a persistent seed bank is one calendar year. Thus, species whose seeds live in or on the soil for <1 year have a transient seed bank, whereas those whose seeds live for ≥1 year have a persistent seed bank. However, dormancy cycling of seeds buried in soil has not been given due consideration in these models. When dormancy cycling is considered, it is shown that seeds of both autumn-germinators and spring-germinators are in the dormant state when they are 1 year old. Thus, unless the seeds live until at least the second germination season (i.e. usually 16–18 months following dispersal), they are, in effect, part of a transient seed bank, having lived through only one germination season. We propose that for seeds of such species to be considered part of a short-term persistent seed bank, they should remain viable and germinable until at least the second germination season, and to be part of a long-term persistent seed bank, until at least the sixth germination season. Our definitions are applicable to seeds with physiological, physical or morphophysiological dormancy, which often require >1 year after maturity to come out of dormancy in nature. We discuss modifications of the seedling emergence method for detection of a soil seed bank, so that they correspond to our definitions of seed-bank strategies.