Towards a Domestication Strategy for Indigenous Fruit Trees in the Tropics - Indigenous Fruit Trees in the Tropics: Domestication, Utilization and Commercialization - page 37

performance trials or directly into compatibility trials. This raises the question of
how the superior mature trees should be identified, especially if it is desirable to
select for multiple traits. This can create a problem, as many traits may be
weakly or negatively correlated (e.g. fruit size and kernel size in S. birrea ;
Leakey, 2005). Consequently, as the number of desirable traits increases, the
number of genotypes superior for all traits diminishes rapidly. Thus, the
selection intensity (and also the number of trees screened) must be substantially
increased, or the expected genetic gain will rapidly decline. For this reason,
only the few most economically important traits (e.g. fruit flesh or nut mass,
taste) should be concentrated on in the early phases of selection.
Two techniques can be used to assist in the identification of superior
mature trees (sometimes called 'elite' or 'plus' trees) producing indigenous
fruits and nuts. The first is to involve indigenous people in the domestication
process and to seek their local knowledge about which trees produce the best
products. Local people usually have good knowledge about the whereabouts of
elite trees, and this knowledge often extends to superiority in a number of
different traits, such as size, flavour and seasonality of production. However,
access to this knowledge has to be earned by the development of trust between
the holder of the knowledge and the potential recipient. Ideally, the recipient
should enter into an agreement that the intellectual property rights of the
holder will be formally (and legally) recognized if a cultivar is developed from
the selected tree. Unfortunately, at present the process of legally recognizing
such cultivars is not well developed and requires considerable improvement.
The second technique for identifying mature elite fruit and nut trees has
recently been extended to a study on marula (Leakey, 2005) following its
development in Cameroon and Nigeria (Atangana et al ., 2002; Leakey et al .,
2002, 2005c). This technique involves the quantitative characterization of
many traits of fruits and kernels, which are associated with size, flavour,
nutritional value, etc. This characterization also determines the extent of the
tree-to-tree variation, which is typically three- to sevenfold, as found in marula
(Leakey et al ., 2005a, b), as well as the frequency distribution, which is
typically normal in wild populations but tends to become skewed in
populations subjected to some selection. The characterization data can then
be used to identify the best combination of traits (the 'ideotype') to meet a
particular market opportunity (for an example see Fig. 2.3). The development
of single-purpose ideotypes (Leakey and Page, 2006) provides a tool for the
development of cultivars with different levels of market focus and
sophistication (Fig. 2.4). In addition to advancing the selection process for
multiple traits, the ideotypes also provide information about opportunities to
select for better partitioning of dry matter between desirable and undesirable
traits. For example, in marula 90% of the dry matter in nuts is typically found
in the shell and only 10% in the valuable kernel, although the range of
variation is 3-16%. The shell : kernel ratio is therefore a trait which could
usefully be included in the ideotype. However, the inclusion of each additional
trait in a multi-trait selection process greatly increases the number of trees that
need to be screened, especially if the traits are independent or only weakly
related.