83(Rejesus et al. 1996 ). Each type of parent material appears to be used more in the
pursuit of certain goals than others. Wild relatives and landraces are more likely to
be used for disease resistance and quality than for yields, while advanced lines and
released cultivars are more frequently used in breeding for yield potential than for
other breeding goals.
Starting from the F 2 generation, the parental traits are combined and become
fi xed in the progeny. The smaller the number of genes determining the expression
of a trait, i.e., the simpler the inheritance, the earlier the generation in which the
progeny plants become homozygous and the sooner selection can be commenced.
The most important plant traits, however, such as yield, quality, biotic resistance,
etc., are polygenic, so selection for these traits is only effective in later generations.
Depending on when plant selection is started, two methods of breeding can be dis-
tinguished, bulk selection and pedigree selection.
When the bulk method is applied, the populations arising from a cross are sown
and harvested for several years without human selection. The essence of this method
is that while the level of homozygosity rises from year to year due to self- fertilisation,
natural selection leads to the increasing frequency of favourable genes. The bulk
method is a cheap way of handling segregating populations in cases where the envi-
ronmental conditions make effi cient natural selection possible for a trait critical for
cultivation. The method is ideal when breeding for tolerance of unfavourable eco-
logical factors, e.g., in locations with a very cold winter. In such areas, plants not
adapted to the environment, i.e., with poor winter hardiness, will die, leading to the
greater frequency of plants with the desired traits. Natural selection may have a
similar success rate in the case of disease epidemics, where the survival and repro-
duction rates of resistant plants will be greater than that of susceptible plants.
The bulk method can be supplemented by mechanical mass selection, which
relies only partially on the selection effect of the natural environment. In addition
the population is artifi cially reduced to increase the ratio of plants with favourable
traits. Selection may be direct or indirect. Direct selection can be performed for
enhancing grain size (Derera and Bhatt 1972 ) and adaptability by selecting for grain
size. An effi cient way of reducing plant height is to cut the population back to the
desired height (Romero and Fey 1966 ). In a special case of mass selection for earli-
ness and drought tolerance, the plant stand is desiccated prior to ripening, after
which selection is made for large grains (Blum et al. 1983 ). After several years of
self-fertilisation, pure line selection is begun in the F 5 –F 6 generation of the bulks,
followed by the testing of selected lines with satisfactory homogeneity.
In the pedigree method, plant selection is commenced in early generations (F 2 or
F 3 ) and is repeated several times until a satisfactory level of homozygosity is
reached. If plant selection is not begun in the F 2 population, a manageable popula-
tion size is generally obtained by manual mass selection. In this case the selected
ears are threshed together and the grains thus obtained are sown for further
selection.
In order for every possible recombination in a segregating population to occur at
least once, a much larger population size would be needed than can be achieved in
practice. Even if a desirable genotype occurred in a population with low frequency,
3 Wheat Breeding: Current Status and Bottlenecks