Agriculture Reference
In-Depth Information
effi cient than direct selection of either produc-
tion trait in early generations (Rebetzke et al.,
2002).
The wheat cultivars Drysdale and Rees were
selected from a BC
2
-derived population devel-
oped from the high-CID recurrent parent
'Hartog'. In side-by-side studies of Drysdale and
Hartog undertaken in about 30 sites in each of two
years, AGT breeding company established an
average yield advantage of 16% and 9% for 2004
and 2005, respectively (Fig. 11.8). The greatest
yield advantage occurred in the drier of the two
years and in the driest of sites.
Evaluation in water-limited environments of
BC
2
-derived, sister lines showed CID to be nega-
tively correlated with aerial biomass and yield
(Rebetzke et al., 2002). In comparisons with the
recurrent parent Hartog, low-CID selections
achieved greater yield through increases in harvest
index and biomass (Fig. 11.19). Kernel size was
the primary yield component affected by selection
for low CID. In contrast, high CID has often been
associated with higher leaf conductance, increased
water use, and growth. Hence positive relation-
ships for CID, biomass, and yield are commonly
obtained in irrigated environments where water
supply was not a major constraint to yield (Fischer
et al., 1998), or in Mediterranean environments
where soil water availability preceding anthesis
was plentiful (Araus et al., 2002). Thus the oppor-
tunity exists to select for high CID to increase
potential yield where water for crop growth is
abundant.
Repeatable genotypic variation has been
reported in wheat for TE (Solomon and Labus-
chagne 2004) and CID (Rebetzke et al., 2002,
2006). These reports emphasize that broad- and
narrow-sense heritability of CID is high when
expressed on a single-plot or line-mean basis.
Further, analysis of mating designs employing
progeny from either F
1
or segregating generations
has shown TE and CID to be under strong addi-
tive genetic control with little evidence for nonad-
ditive gene action (Solomon and Labuschagne
2004; Rebetzke et al., 2006). Together, these
results indicate that family selection for altered
TE and/or CID in early generations will maintain
altered CID with inbreeding.
Few QTL analyses have been reported for CID
in wheat. Genotypic distributions for CID are
commonly gaussian with evidence for transgres-
sive segregation (Rebetzke et al., 2008a). QTL
studies have been undertaken for wheat at CSIRO
for multiple populations across different environ-
ments (Table 11.2). These indicate a large number
of genomic regions of small effect, many of which
are common across populations (Rebetzke et al.,
2008a). Some of these regions collocate with
regions for plant development and plant height, as
well regions for the stomatal-related traits, canopy
temperature, stomatal conductance, and leaf chlo-
rophyll content (Fig. 11.20; Table 11.2). Reduced
CID is commonly associated with later fl owering,
highlighting the need for caution when relating
CID to yield in a terminal drought study of popu-
lations in which fl owering is not carefully moni-
tored (Sayre et al., 1995; Condon et al., 2002).
Similarly, lower CID is associated with reduced
plant height and, particularly, the presence of
Rht-
B1b
and
Rht-D1b
dwarfi ng genes (Fig. 11.20).
A number of the CID QTLs identifi ed for leaf
tissue collocated with QTLs for CID measured on
mature grain (Fig 11.20). Yet despite this com-
monality, alleles for increased CID are sometimes
not consistent for leaf and grain CID QTLs. For
example, parents contributing low leaf-tissue CID
alleles at the
Ppd
,
Rht-B1
, and
Rht-D1
loci
6
5
4
3
2
1
0
Grain
yield
Total
biomass
Grain
weight
Grain
number
Harvest
index
Fig. 11.19
Change in grain yield and yield components with
selection for reduced carbon isotope discrimination in BC
2
-
derived lines developed from recurrent parent 'Hartog'
evaluated in nine environments.
