Environmental Engineering Reference
In-Depth Information
when many genotypes are intercrossed, in which case seeds are gathered from the inflorescence of
all parents involved, which prevents the identification of the pollen source; and (3) free pollination,
in which seeds are harvested from plants growing freely.
Crossings should be planned in such a way as to maximize the probability of genotype selection
that can be released as commercial cultivars. Thus, an alternative that is greatly used is to choose
parents with good economically interesting performance traits (Matsuoka et al. 1999a), which natu-
rally occurs for commercially used cultivars. It is noteworthy that this can lead to a narrowing of
the genetic basis (Lima et al. 2002).
Because sugarcane is a type of allogamy, the crossings should be planned so as to avoid the
occurrence of interbreeding between relatives. This can be achieved based on the genealogies
of materials as well as genetic divergence obtained with molecular markers (Lima et al. 2001).
Other criteria should also be used such as trait complementarity; ability to combine material
during the crossings; and the capacity for each material to produce good populations throughout
time.
Some authors mention the use of estimates of the genetic parameter named heritability that in
its restricted form takes into account the sexual phase. This can be considered in the prediction of
crosses, because heritability indicates how the interest traits are transmitted to the offspring. Some
results indicate that the predominant gene action in the Brix content is additive, whereas for the
other components the output is not additive (Hogarth 1980; Wu et al. 1980). Hogarth (1977) cites
that for stem production, the dominant variance has shown the same magnitude as the additive,
where the epistatic variance is predominant for stem weight. Hogarth (1987) showed high heritabil-
ity values for the fiber content trait. For volume and number of stems, the dominant variance proved
to be important (Hogarth et al. 1981). Bressiani (1993) reported high heritability values for length
of stem and Brix in Brazilian conditions. For rust, the heritability has been high (Hogarth et al.
1983; Bressiani and Sanguino 1994), which also usually happens with other diseases (Matsuoka
et al. 1999a), indicating that the selection of parents may be resistant effective. However, there are
examples of transgressive segregation.
21.6.2 i
nitial
S
tagES
of
S
ElEction
There are many variations on how the breeding programs develop this stage, especially in terms
of selection rates, group size, number of locations, and replications (Simmonds 1979; Skinner et
al. 1987; Matsuoka et al. 1999a). In general, a low intensity selection is applied in the early phases,
selecting only for traits with high heritability. This is followed by an increase in selection intensity
as the experimental precision increases, and cultivation recommendation only when there are many
experimental results in different places and years of cultivation.
Broad sense heritability, which should be considered in these phases, has several different val-
ues reported in literature. For example, in relation to vigor, some studies indicate low heritability,
although several breeding programs have obtained good results with intense selection for this trait
(Skinner et al. 1987). These authors report different heritability values for various traits, such as
sugarcane yield; Brix yield per hectare; number and diameter of stems; and resistance to rust and
smut. These results suggest that in early stages selection is more effective for Brix and for resistance
to rust and smut, but these data should be interpreted with caution given the low number of available
estimates (Matsuoka et al. 1999a).
Another very important feature to be considered in these phases is the genotype regrowth ability.
Despite the small number of available estimates in this respect, many programs consider this aspect
in the selection process (Giamalva et al. 1967; Watkins 1967; Giroday 1977; Mariotti 1977; Bond
1978; Skinner et al. 1987; Arizona 1994; Matsuoka et al. 1999a), which can lead to many genotypes
in the early phases being discarded.
Several selection strategies for these steps have been used and are briefly described below.
