B. Mechanisms by Which Genotypes Differ in TE
Any factor that influences genetic variation in either gsorAin a disproportionate manner would influence
and thus TE [38]. If variation in Awas the only cause of variation in Pi, increasing photosynthetic ca-
pacity should lower Pi/Paand therefore lower . In this situation, TE would increase and the relationship
betweenand plant biomass should be negative [124]. In groundnut, differences in Aare reported to be
largely responsible for TE variation, as dry matter production is negatively correlated with in pots
[10,74] and at the canopy level [66,96]. Significant variation in Aper unit leaf area has been reported in
groundnut genotypes and there is also heterosis for this trait [68,125–127]. In cowpea, genotypic means
for TE were positively correlated with Abut only weakly correlated with gs, indicating that genotypic dif-
ferences in TE were mainly due to differences in A[112]. Similarly, in sunflower, tomato, and wheat
genotypic differences in TE were due to differences in A[128,129].
A strong positive correlation has been observed between and specific leaf area (SLA) among
groundnut genotypes [101,130,131]. This is consistent with the foregoing hypothesis that high TE geno-
types have higher A. Indeed, the genotypes with thicker leaves (low SLA) had significantly higher leaf ni-
trogen contents, again indicative of higher photosynthetic capacity. The significant application of these
observations is that breeders could use the inexpensively measured SLA, in lieu of , to screen for high
TE among groundnut genotypes within specific environments [74].
However, if gsis the main source of variation in Pi/Pa, greater gsshould increase Pi/Paand therefore
increase. In adequately irrigated coffee, higher TE values of some of the genotypes tested were associ-
ated with reduced stomatal aperture rather than increased Aat a given gs[99]. This suggests that high TE
may restrict yield when water supply is not limiting. Thus, in this case, as in wheat, selection for higher
could lead to increased biomass production but with decreased TE [132]. For example, in crested
wheatgrass, greater TE in low clones resulted from a proportionately greater decline in gsthan in A
[106]. Similar results were reported for chickpea [107]. However, variation in Pi/Paamong wheat geno-
types is approximately equal to variation in leaf gsand in A[65,133–135]. In wheat, it was reported that
gscovaried with A, with the change in gsbeing relatively greater [135]. This means that there could be a
positive correlation between AandPi/Pa. The effect of this on growth may be compounded if genotypes
with large Pi/Papartition more carbon into shoots [136].
Cultivar differences in may also result indirectly from genetic variation in root characteristics af-
fecting the level of water stress experienced by the canopy [98,137]. Differences in root growth affect the
degree of dehydration postponement, and this could prolong gas exchange activity and the maintenance
of relatively high Piand thus [137].
C. Genetic Variation and Genetics of TE and
Genetic variation in TE and has been reported in wheat [65,124,132,138], barley [95], tomato [86],
sunflower [100], chickpea [107], groundnut [64,66,74,131], cowpea [88], alfalfa [139], and coffee [99].
In wheat, variation in among genotypes is typically around 2 10 ^3 [138]. This is equivalent to a
variation in TE of 59% [138]. In groundnut, genotypic variation in TE is estimated as about 65% [64].
Based on extreme cases of genotypes that differ in TE, it was reported that cowpea genotypes such as
vita 7 and 8049 had nearly 67% higher TE values than those of other genotypes tested [111]. Also, ear-
liness is generally associated with low TE in cowpea; however, significant genotypic differences were
noticed within any given maturity group, suggesting that these two traits are not necessarily linked
[111]. Similarly, tall landrace genotypes of wheat, which are also late maturing, had higher TE than did
the modern dwarf and semidwarf genotypes [124]. However, among Australian wheats, low values of
and thus high TE have been found to be strongly associated with the WW15 genetic background,
which was introduced into Australia from CIMMYT as a major source of the dwarfing gene in Aus-
tralian wheat.
The utility of a trait for selection in plant breeding programs is strongly enhanced by the consis-
tency of genotypic ranking across environments [112]. Based on studies with wheat, cowpea, crested
wheat grass, groundnut, and beans, it was found that genotypic ranking for across environments is
consistent [37,38,101,109,111,112,132,138,140]. For crops such as groundnut, it was shown that geno-
typic ranking for was maintained during ontogeny [74] (Figure 3). However, in crops such as wheat,
genotypic ranking could change between the early vegetative stage and the heading and grain filling
844 SUBBARAO AND JOHANSEN