Three types of markers-isozymes, RAPDs (random amplified polymorphic DNAs), and wheat microsatellites- were tested on wheat, spelt, and four wild wheat relatives (Aegilops cylindrica, Elymus caninus, Hordeum marinum, and Agropyron junceum). The aim was to evaluate their capability to provide specific markers for differentiation of the cultivated and wild species. The markers were set up for subsequent detection of hybrids and introgression of wheat DNA into wild relatives. All markers allowed differentiation of the cultivated from the wild species. Wheat microsatellites were not amplified in all the wild relatives, whereas RAPDs and isozymes exhibited polymorphism for all species. The dendrograms obtained with RAPD and isozyme data separated Swiss wheat cultivars from those collected in Austria and England, while no difference was found between Swiss spelt and wheat. RAPD data provided a weak discrimination between English and Austrian E. caninus. The microsatellite-based dendrogram discriminated populations of Ae. cylindrica, but no clear separation of H. marinum from E. caninus was revealed. The similarity matrices based on the three different sets of data were strongly correlated. The highest value was recorded between the matrices based on RAPDs and isozymes (Mantel's test, r = 0.93). Correlations between the similarity matrix based on microsatellites and matrices based on RAPDs and isozymes were lower: 0.74 and 0.68, respectively. While microsatellites are very useful for comparisons of closely related accessions, they are less suitable for studies involving less-related taxa. Isozymes provide interesting markers for species differentiation, but their use seems less appropriate for studies of within-species genetic variation. RAPDs can produce a large set of markers, which can be used for the evaluation of both between- and within-species genetic variation, more rapidly and easily than isozymes and microsatellites.