Seed losses by surface wash in degraded Mediterranean environments

The susceptibility of seeds of 10 species to be removed by water erosion from an abandoned field and a badland site in SE Spain was determined. Rainfall simulation experiments were conducted over 0.24 m(2) plots for a 22 minute duration, with an intensity of 55 mm . h(-1). A mixture of 24 seeds of each species had been placed in these plots, and the number of seeds lost were measured. Seed losses were analyzed in relation to soil erosion rate and overland flow and to seed parameters such as seed size and form. Seed losses were low in all cases. Total seed losses were less than 10% for all replicates, and there was no species that lost more than 25% of its seeds in any replication. There was a strong relation between total seed losses and runoff in the badland site, but not in the abandoned field. Seed losses were not significantly related to soil losses in either site. There was a wide variation of seed losses between species, but this variation could not be explained by seed weight or shape. Other seed features such as mucilage secretion and the presence of appendices, as hairs or wings, may explain the differences in seed losses at species level. (C) 1997 Elsevier Science B.V.