In 1992, Vaillancourt and Weeden discovered a very important mutation for studying cowpea evolution and domestication. A loss of a BamHI restriction site in chloroplast DNA characterized all domesticated accessions and a few wild (Vigna unguiculata ssp. unguiculata var. spontanea) accessions. In order to screen a larger number of accessions, primers were designed to check this mutation using PCR RFLP or direct PCR methods. Using these new primers, 54 domesticated cowpea accessions and 130 accessions from the wild progenitor were screened. The absence of haplotype 0 was confirmed within domesticated accessions, including primitive landraces from cultivar-groups Biflora and Textilis, suggesting that this mutation occurred prior to domestication. However, 40 var. spontanea accessions distributed from Senegal to Tanzania and South Africa showed haplotype 1. Whereas this marker could not be used to identify a precise center of origin, it did highlight the widely distributed cowpea crop-weed complex. Its very high frequency in West Africa could be interpreted as a result of either genetic swamping of the wild/weedy gene pool by the domesticated cowpea gene pool or as the result of domestication by ethnic groups focusing primarily on cowpea as fodder.