Post-dispersal seed anchorage to soil in semiarid plant communities, a test of the hypothesis of Ellner and Shmida

In this paper we test the Ellner and Shmida's hypothesis that in semiarid environments aridity may select for the lack of seed dispersal mechanisms (atelechory) of many plants, whereas post-dispersal selective forces such as soil erosion, seed predation or limitations to water uptake by seeds may select for structures facilitating seed anchorage to the ground (antitelechory). We analyzed the proportions of species with seed anchorage mechanisms and that of atelechoric species in shrublands colonizing flat areas and hillslopes in two sites differing in climate dryness. Their relation with several soil properties involved in runoff generation, seed-soil contact and water uptake by seeds and with nest density of granivorous ants was also explored. Our results support the hypothesis that in semiarid shrublands the proportion of species with seed anchorage mechanisms increases because of soil erosion but not because of climate dryness. This is the first time that a direct relation between the proportion of species with seed anchorage mechanisms and soil erosion is shown in plant communities; supporting the view that soil erosion shapes species composition in communities.