The seed bank dynamics of annuals in two Mediterranean temporary marsh
es were studied in order to investigate the impact of the disturbance
regime on coexistence. The disturbance regime is made up of drought pe
riods in the marshes and of predation. The density of diaspores buried
in the sediment was measured in 4 consecutive years together with the
biomass produced during two growing seasons. The total density of dia
spores in the sediment varied markedly from year to year but always re
mained at a high level for all species (between 73000 and 800000 per m
(2)). These perennial seed banks allowed the populations to be maintai
ned despite two successive years without any reproduction. In experime
ntal conditions, the seed bank was not depleted after the sediment sam
ples had been submerged during 5 successive periods. The majority of t
he seedlings emerged as soon as the first period of flooding and tempo
rary droughts stimulated the germination of Zannichellia spp. and Ranu
nculus baudotii. The decrease in the number of seeds buried in the soi
l was essentially due to germination. This was stronger when the seeds
were recently produced and situated at the surface of the sediment. T
he seed density varied only slightly with the depth of the marsh, exce
pt in the case of Ruppia maritima. The seed bank is highly variable on
a spatial microscale bur no aggregation pattern could be demonstrated
. In the field, the biomass of a population generally appeared not to
be limited by the availability of the diaspores, notably because of th
e ability to propagate vegetatively which is shown by all species stud
ied. Similarly, a high vegetative biomass does not guarantee a high re
productive output as the droughts occur in a density-independent way.
The success of a species is related to the interaction between its lif
e history traits and the environmental conditions met each year. The m
acrophyte communities we studied function in non-equilibrium, with pat
terns close to those described in lottery models of coexistence. Estab
lishment, growth and reproduction fluctuate strongly from one year to
another for all species, in a partially asynchronous way. This would e
nable competitive exclusion to be indefinitely delayed. The seed bank
introduces a storage effect by magnifying the effect of the ''favourab
le years'' in comparison with the ''bad years''.