Agriculture Reference
In-Depth Information
sugarcane. Early generation selection in potatoes, see
Figure 7.8 has been shown to be at best a random reduc-
tion in the number of genotypes in the breeding scheme.
There has been some evidence that selection in the early
generations was producing an undesirable response and
that selections were not always the genotypic lines most
suited to agricultural conditions.
Significant correlations have been found between
sugar cane seedlings and later clonal generations for
stalk and stalk diameter. These associations although
statistically significant would result in large selection
Type I errors and selection ratios less than 2.0. With
such results it may be difficult to justify the expense
and effort that such selection would involve.
Early generation selection of grasses using small plots
resulted in identifying lines which did not perform
well under sward conditions where inter-plant compe-
tition was greater. A similar response has been noted in
potato where selected lines were less competitive under
field stand conditions due to selection being carried out
under wide plant spacing.
Despite the relative inefficiency of selection in the
early generation stages most plant breeding schemes
usually discard by far the greatest proportion of genetic
variation in the first and second rounds of selection. It is
certainly not uncommon to have cases where 99.9% of
genotypes are discarded in the first or second selection
stage and that this has been achieved using small plots
and without replication.
In summary, selection in the early generations is
usually affected by:
allows breeders to say that it is best to select for this
now and at this intensity. There are some simple ques-
tions which can help in making these decisions. These
include:
•
What are my breeding objectives and what characters
are to be included throughout the
whole selection
process
?
•
What characters can be most easily and most eco-
nomically assessed on small plots with minimal
replication?
Which characters are most heritable (i.e. high
h
2
)
•
?
Which have low heritability? Selection in the early
generation should be based on the most heritable
traits
•
Which characters have highest priority? For example
which characters
must
a new cultivar have? Either by
being important (i.e. high yield or specific quality) or
by legislation (i.e. low glucosinolates and erucic acid
content in canola)
Never forget the golden rule of any selection, that
a breeding scheme should never carry more individual
genotypes than can be
efficiently screened
. It is almost
always more effective to evaluate fewer lines with greater
accuracy than to use an ineffective selection scheme.
Intermediate generation selection
It is assumed that at the intermediate generation selec-
tion stage, the large initial population (usually thou-
sands) has been reduced to a practical number which
will allow more detailed assessment (usually hundreds
of lines). It is also assumed that by this stage there has
been a simultaneous increase in the availability of plant-
ing material. As a result of fewer lines and more planting
material, it is possible to organize evaluation trials which
have reasonable plot sizes (may differ according to the
crop), and replication of all test entries is possible.
The number of lines that require testing at the inter-
mediate stage will still be large enough to dictate that
evaluation is still restricted to only one (sometimes two)
locations.
•
Limited amounts of planting material
so it is not
possible to have sophisticated experiments involving
large plots, high replication and multiple sites
Large numbers of genotypes
need to be evaluated
which also usually results in small plots (often single
plants), low levels of replication and single location
trials
•
As a result, many of the initial evaluations are carried
out by visual inspection rather than, say, actual record-
ing yield. Similarly many of the more “difficult to assess”
traits including polygenic disease or pest resistance or
quality character cannot be easily taken into account.
It is difficult to determine exactly what charac-
ters are to have priority in early generation selection.
Unfortunately there is not any simple equation which
Field trials
Field testing is a major part of all selection, and inter-
mediate selection is no exception. Test entries should
