Ecological correlates of seed persistence in soil in the north-west European flora

I Using the data in a recently published seed bank database for north-west Europe, we describe how a species' seed bank behaviour can be characterized by a single 'longevity index', and investigate how representative the information in the database is of the north-west European flora. We also test the hypotheses that seeds of short-lived species are more persistent than those of long-lived species, and long-lived seeds are characteristic of species living in disturbed habitats. 2 The data are not representative of the north-west European flora as a whole; they are a fair reflection of a research effort that has been largely directed towards grassland and arable weeds. Data for rare species, non-agricultural habitats in general and wetlands, rocky habitats and native woodland in particular, are scarce or absent. 3 Annuals and biennials almost always have more persistent seeds than related perennials, and this difference is most striking when, as in Anagallis arvensis and Aphanes arvensis, the short-lived species have moved well away from the 'core' habitat of the family. Confamilial monocarps and polycarps do not differ consistently in seed mass. 4 Gradients of habitat disturbance are accompanied by predictable changes in seed persistence, which are themselves often (but not always) accompanied by parallel shifts in seed size. These results suggest that increasing habitat disturbance (i.e. increasing density-independent mortality) always selects for increased seed persistence, confirming both theory and previous analyses. However, increased seed persistence is not always associated with reduced seed size. This is because persistence depends not only on seed size, but on other traits, many of them physiological. In many habitats the probability of seed burial is strongly linked to seed size and shape, but in arable habitats cultural practices have broken this link.