The range of genetic resources available for the improvement of grain legumes varies greatly in both its extent and accessibility. This can be related to the biosystematic relationships and the geographic dispersion of the crops together with the evolutionary age of the taxa from which they arose and related taxa. Harlan and de Wet’s gene pool system can usefully be applied to the legumes. Primary gene pools are often extensive, secondary gene pools are usually restricted or non-existent. Tertiary gene pools, though extensive, would require development of completely new techniques for exploitation.