In general, the formal seed system handles only varieties produced by systematic plant breeding in public or private research stations.  Such varieties have a defined origin and are subject to official testing before being registered in a national list, which makes them eligible for multiplication and certification. The precise relationship between testing, registration, release and certification varies somewhat between countries but the general principles are similar. With increasing interest in traditional varieties, the question which arises is whether they can have any place in an official variety control system. 
        The easy answer is 'NO' - because the testing system is geared to receiving new varieties from breeders.  Furthermore, traditional varieties may be more variable and may therefore fail in a conventional DUS testing system. However, this rigid policy may deprive farmers of varieties which could be useful and it may accelerate the loss of valuable well-adapted material because of its uncertain quality status and identity.  It is therefore appropriate to consider how the formal seed system can accommodate traditional varieties and perhaps help to upgrade them.
        The first problem is to agree on an identity - since traditional varieties are likely to be more variable and less precisely defined. As a first step, a number of samples of the 'variety' should be collected from reputable sources, if possible in the area with which it most closely associated. Sometimes there is a strong link between a variety and a particular location.  If these samples are grown in small plots and inspected by a group of well-informed farmers and technical staff it should be possible to agree on which type most clearly represents the known characteristics of the variety. This type can then be described according to the standard variety descriptors for the crop. The same process could be repeated in a subsequent year if further seed sources are identified, but the concern should be to agree as soon as possible on the representative type.
        The next stage is to produce a nucleus stock of the material. This can be done by growing spaced plants or small plots and closely inspecting them to remove any widely divergent forms. The intention is not to eliminate variation, but rather to reduce the variability to an acceptable level which will facilitate subsequent management of the variety. Knowledgeable farmers should again be involved in this process to obtain the benefit of their experience. One or two multiplications should be carried out under close supervision to assess the stability of the material and eliminate any obvious contaminants. At this stage the registration and certification authority should be asked to inspect the plots to obtain their agreement about the status of the variety and of the nucleus stock as a basis for further multiplication. They may certify a seed crop as 'pre-basic' seed.
        Finally, it is necessary to find a maintainer who will be responsible for the nucleus material. This could be a research station or a group of farmers who have experience of growing and using the variety.  Forage varieties may be associated with a particular management system which contributes to their merits.  This system should be maintained as far as possible in the maintenance and early generations to preserve these particular characteristics. Once a maintainer has been identified, the certification authority should confirm the identity and status of a nucleus stock and the procedures for maintaining it.
        Multiplication and certification can then be initiated as for any other variety. The seed quality standards applied will be identical to those for any other variety of the species, since there is no reason why those standards should be any lower in a local variety. The genetic purity standards may need slightly greater flexibility if the variety is recognised as being somewhat variable.   It may be helpful for the registration and certification authority to create a recognised category of 'local varieties' for which slightly different genetic criteria are applied.  This may be particularly useful if the variety is being grown for use in a particular area for which it is well adapted thus avoiding the need for multi-location testing.
        If the variety contains some variability, then selection may continue in subsequent cycles of 'maintenance' but it is important not to 'over-select' simply to achieve uniformity while risking the loss of adaptability. On the other hand, it may be possible to apply a strong selection pressure to identify superior lines within the population - in which case these may be given a separate name and registered in the normal way as a distinct variety.
        This approach is not only applicable to developing countries. There are examples from Europe of traditional varieties which have entered national lists and certification schemes without any difficulty such as 'Scots Timothy' and 'Kent Wild White Clover', both were successful forage crop varieties in the UK for many years. In some cases, a designated group of farmers may become the maintainers and primary seed producers, thereby gaining an interest in maintaining the quality and identity of the variety and protecting it against contamination. However, legal protection of such varieties is best avoided because of the complication of deciding on the ownership.  In the USA there is also a facility for certifying the origin of the material if it is not fully registered as a variety. Michael Turner, Seed Unit, ICARDA; E-mail: M.Turner@cgiar.org