Neighbor competition modifies demographic responses of desert dune annuals to gradients in sand stability

A field experiment was conducted to test the hypothesis that competitive interactions affect demographic responses of desert dune annuals to gradients in sand stability. The experiment was based on a three-way factorial design with two types of neighbor treatments (complete neighbor removalvs. control), two types of habitats (stablevs. unstable sands) and two species which characterize sandy areas of the western Negev Desert (Neurada procumbensandBromus fasciculatus). Each of the eight treatment/habitat/species combinations was characterized by three demographic variables: seedling survival to reproduction, fruit number per reproductive plant and fruit number per seedling. All of these variables exhibited significant interactions between the effects of neighbor removal and habitat type, indicating that neighbor conditions were important in determining the observed patterns of demographic variation. Moreover, while plants ofNeurada procumbensgrowing without neighbors suffered more mortality on unstable sands, conspecific plants growing at natural density conditions suffered more mortality on stable sands. In the case ofBromus fasciculatus, individuals growing in the absence of competitive effects produced more fruits on stable sands, but conspecific plants growing at natural density conditions exhibited higher fruit production on unstable sands. These findings indicate that competitive effects not only modify, but may even reverse demographic responses of desert dune annuals to gradients in sand stability. Based on these results and evidence from two other studies it is suggested that seed dispersal from abiotically favorable into less favorable habitats may increase fitness of desert annuals by reducing the average effect of competition