Agriculture Reference
In-Depth Information
Cultivar release
The last stage before releasing a line or dropping
it from the program (the more common event) is
extensive fi eld evaluation (Color Plate 29). The
generation when single-plant selection is termi-
nated depends upon the level of homogeneity
desired in the released cultivar. Heterogeneity in
a released cultivar is a direct result of the genetic
differences between the parents and the fi nal gen-
eration of single-plant selection (Baenziger et al.,
2006b). In the preceding example, with 10 segre-
gating loci and selection in the F
6
generation,
73% of the lines were homozygous at all loci.
However, 27% of the lines were heterozygous at
one or more loci, and segregation of these hetero-
zygous loci with continued selfi ng will lead to
heterogeneity in later generations. If plant selec-
tion ends in the F
4
generation, only 26% of the
lines will be homozygous at all 10 loci and 74%
of the lines will be heterozygous at one or more
loci, meaning that they will produce heteroge-
neous lines in future generations. Heterogeneity
for traits not related to morphological appearance
or value for cultivation and use may be tolerated
in a fi nished cultivar, such as variation in some
gliadin subunits. Where heterogeneity is tolerated
(e.g., in many parts of the US), within-line selec-
tion often ends in the F
4
generation. However, in
Europe with more stringent uniformity (homoge-
neity) standards, fi nal single-plant selection is
delayed until later generations (F
6
or later).
environments where it may eventually be grown
and also in those environments where it should
not be grown, because growers should know
which cultivars may or may not be grown on their
farms. Regional performance trials are usually
very expensive, because they must be located on
and away from the main breeding nurseries to
adequately represent the target environments
(Ortiz et al., 2007).
Resource allocation is extremely important in
this phase of wheat breeding, and numerous
researchers have studied the optimal numbers of
replications, locations, and years to estimate line
performance (Comstock and Moll 1963; Bos 1983;
Carter et al., 1983; Crossa 1990; Gauch and Zobel
1996; Cullis et al., 1996a,b, 2000; Yau 1997). The
inference space from the dataset generated pro-
vides the basis for predicting a cultivar's future
performance. The importance of having the most
effi cient regional performance assessment strat-
egy is that the cost per plot is relatively similar,
though it will change with more locations; hence
available resources will largely determine the total
number of plots that can be used. With a fi xed
number of plots, determining how to divide them
among replications and locations within a year is
critical, as is the determiniation of the most effi -
cent experimental design and subsequent statisti-
cal analysis (Smith et al., 2005).
Though not unique to wheat, it should be
understood that with the exception of wheat pro-
duced for irrigated production, rarely are rainfed
nurseries consistently devoid of spatial variation.
Hence methods that can remove spatial variation
are necessary (Stroup et al., 1994). Removing
spatial variation requires augmented or replicated
designs. Single-replication trials which are popu-
lar in some crops are used in wheat only where
seed is limited.
If the variation among locations is expected to
be similar to the variation among years, then it is
possible to substitute locations for years. Knowing
the number of years that lines need to be evalu-
ated is important to provide the grower with suf-
fi cient data over environments that they can make
educated decisions on cultivar selection (Eskridge
1990). To quickly release a line, breeders usually
try to substitute locations for years with the hope
Understanding the phenotype
All plant breeders understand that the measured
phenotype is a function of the genotype, the envi-
ronment in which the genotype is grown, and the
G
E interaction (Acquaah 2007). An inbred line
developed or selected only at one location—a
single environment—is likely to exhibit a narrow
or specifi c adaptation to that environment. This
approach is like developing a line in a greenhouse
and then expecting it to be adapted across a broad
range of environmental conditions. Hence a pre-
requisite for release of a cultivar is that its pheno-
type must be thoroughly understood. In evaluating
an experimental line it needs to be tested in those
×
