RAPD divergence caused by microsite edaphic selection in wild barley.

Random amplified polymorphic DNA polymerase chain reaction (RAPDPCR) was used to assess genetic diversity in four subpopulations (86 individuals) of wild barley, Hordeum spontaneum, sampled from Tabigha microsite near the Sea of Galilee, Israel. The microsite consists of two 100 m transects that are topographically separated by 100 m, each equally subdivided into 50 M of basalt and terra rossa soil types. Despite the same macroclimate characterizing the area around the Sea of Galilee, the microsite offers two edaphically different microhabitats, with basalt being a more ecologically heterogeneous and broaderniche than the relatively drier but more homogeneous and narrowniche terra rossa. Analysis of 118 putative loci revealed significant (P<0.05) genetic differentiation in polymorphism (P0.05) between the two soils across the transects with P being higher in the more heterogeneous basalt (mean P0.05 = 0.902), than in terra rossa (mean P0.05 = 0.820). Gene diversity (He) was higher in basalt (mean He=0.371), than in terra rossa (mean He=0.259). Furthermore, unique alleles were confined to one soil type, either in one or both transects. Rare alleles were observed more frequently in terra rossa than basalt, and in transect II only. Gametic phase disequilibria showed a larger multilocus association of alleles in basalt than terra rossa, and in transect I than II. Spearman rank correlation (r(s)) revealed a strong association between specific loci and soil types, and transects. Also, analysis of multilocus organization revealed soilspecific multilocusgenotypes. Therefore, our results suggest an edaphically differentiated genetic structure, which corroborates the niche widthvariation hypothesis, and can be explained, in part, by natural selection. This pattern of RAPD diversity is in agreement with allozyme and hordein protein diversities in the same subpopulations studied previously.