• 1In recent decades there has developed a very general appreciation of the need to build, maintain and use collections of crop plants with the primary object of sustaining genetic advance by plant breeding on into the indefinite future. The fact of genetic erosion is universally acknowledged and substantial advances in some practical aspects of genetic resource conservation work have been made. Many good collections have been assembled, though some crops have been poorly served, even ignored, and maintenance, study and utilization of those that have been collected have often left much to be desired.
  • 2This review is devoted to the most neglected aspect of all, namely utilization. Traditionally, collections have been regarded as sources of ‘genes’, usually disease resistances, to be exploited by backcrossing (Introgression) into adapted stocks. Thousands of such backcrosses have been made, with very varied success; there have been successes but also many failures, due usually to the weakness of ‘vertical resistances’ as protection against adaptable pathogens. By contrast, the need to broaden the genetic bases of many crops, far beyond the confines possible for Introgression programmes, has often been recognized but rarely explored. Thus a distinction is drawn between Introgression and Incorporation as the two fundamental methods of using crop collections (see Fig. 1).
  • 3Incorporation implies the systematic exploitation of a large array of genetic variability in such a way as to generate a mass of newly adapted stocks usable as parents in breeding programmes. Genetic principles are simple and obvious, always based on recurrent cycles of recombination and mass selection. Progress is likely to be slow (indeed usually must be slow), a fact which has often been a discouragement to such programmes in the past.
  • 4Examples of long-term, systematic Incorporation programmes are few, and the more important ones are summarized. They relate topotatoes, sugar cane and maize, which are well advanced; and cocoa, oil palm and (prospectively) rubber, which are well begun. Time scales depend upon the biology of the crop and methods; typically, a few decades are necessary to make serious progress.
  • 5In discussion, the following points are emphasizedthe essential nature of Incorporation (= Base broadening) in all well-bred crops in the longer term and into the indefinite future; the weakness of phenotypic assessment of genetic potential (usually referred to as ‘evaluation’); the fact that Incorporation is esentially simple, even if slow; the nomenclatural confusion that surrounds the terms Genetic Enhancement, Prebreeding, Genetic Vulnerability; the fact that the common stereotype of a crop collection as a collection of seed packets is often wrong because many crops are clonal and/or have short-lived, unstorable seeds; and the socio-politico-bureaucratic complications that arise because of the current collapse of publicly supported research, the associated short-term ideology, and the fact that serious genetic resource work, including Incorporation programmes, demands long-term commitment to the interests of our successors, not just to ourselves.