A recent approach to detecting genetic polymorphism involves the amplification of genomic DNA using single primers of arbitrary sequence. When separated electrophoretically in agarose gels, the amplification products give banding patterns that can be scored for genetic variation. The objective of this research was to apply these techniques to cultivated peanut (Arachis hypogaea L.) and related wild species to determine whether such an approach would be feasible for the construction of a genetic linkage map in peanut or for systematic studies of the genus. Two peanut cultivars, 25 unadapted germplasm lines of A. hypogaea, the wild allotetraploid progenitor of cultivated peanut (A. monticola), A. glabrata (a tetraploid species from section Rhizomatosae), and 29 diploid wild species of Arachis were evaluated for variability using primers of arbitrary sequence to amplify segments of genomic DNA. No variation in banding pattern was observed among the cultivars and germplasm lines of A. hypogaea, whereas the wild Arachis species were uniquely identified with most primers tested. Bands were scored (+/-) in the wild species and the PAUP computer program for phylogenetic analysis and the HyperRFLP program for genetic distance analysis were used to generate dendrograms showing genetic relationships among the diploid Arachis species evaluated. The two analyses produced nearly identical dendrograms of species relationships. In addition, approximately 100 F2 progeny from each of two interspecific crosses were evaluated for segregation of banding patterns. Although normal segregation was observed among the F2 progeny from both crosses, banding patterns were quite complex and undesirable for use in genetic mapping. The dominant behavior of the markers prevented the differentiation of heterozygotes from homozygotes with certainty, limiting the usefulness of arbitrary primer amplification products as markers in the construction of a genetic linkage map in peanut.