Ever since woody perennials were discovered, or rather re-discovered, as
valuable assets of land use in general and agroforestry systems in particular,
many lists of species considered useful have been circulating. The number of
species qualifying as potentially promising in an agroforestry context runs up
to several thousands. Burley (1984) mentions not less than 23 lists containing
more than 1000 species, and these are not exclusive of others. The palm
family (Palmae) alone comprises at least another 1700 different species
representing more than 200 genera (Johnson, 1983), many of which also have
great potential for their use in agroforestry.
Given the fact that it is only about 15 years ago when the international
community of scientists started to appreciate and study agroforestry as a new
approach to land-use, it is not surprising that there are considerable limitations
in the amount of qualitative and quantitative knowledge. This is particularly
true for the majority of those woody perennials considered as potentially
promising candidates for agroforestry. Aspects important to their productive
and service functions in agroforestry systems, such as phenology (providing
vital information on flowering and fruiting times) or morphology (indicating
height and canopy development and rooting pattern), often prove to be
unknown. Even questions on such important tree features as deciduousness
or whether a tree is monoecious or dioecious remain unanswered in many
cases.
What is admittedly an unsatisfactory state of affairs with regard to qualitative
knowledge proves to be even worse in quantitative terms. Data on
direct productivity of trees and shrubs, such as annual fruit, leaf, or pod
production per mature tree are rare and often difficult to retrieve from
literature in which they may be contained.