BORRELL ET AL.

: STAY-GREEN AND DRY MATTER PRODUCTION IN SORGHUM 1037

nation to temperature amongst diverse sorghum hybrids. Field Wolfe, D.W., D.W. Henderson, T.C. Hsiao, and A. Alvino. 1988a.
Crops Res. 31:295–308. Interactive water and nitrogen effects on senescence of maize. I.
Waggoner, P.E., and R.D. Berger. 1987. Defoliation, disease and Leaf area duration. Agron. J. 80:859–864.
growth. Phytopathology 77:393–398. Wolfe, D.W., D.W. Henderson, T.C. Hsiao, and A. Alvino. 1988b.
Wanous, M.K., F.R. Miller, and D.T. Rosenow. 1991. Evaluation Interactive water and nitrogen effects on senescence of maize. II.
of visual rating scales for green leaf retention in sorghum. Crop Photosynthetic decline and longevity of individual leaves. Agron.
Sci. 31:1691–1694. J. 80:865–870.

Does Maintaining Green Leaf Area in Sorghum Improve Yield under Drought? II.
Dry Matter Production and Yield
Andrew K. Borrell,* Graeme L. Hammer, and Robert G. Henzell

ABSTRACT high frequency of this water limitation in Australian

Retention of green leaf area at maturity (GLAM), known as stay- sorghum-growing environments.
green, is used as an indicator of postanthesis drought resistance in A mechanism of resistance, known as stay-green (Ro-
sorghum [Sorghum bicolor (L.) Moench] breeding programs in the senow, 1977), is indicated by maintenance of green
USA and Australia. The critical issue is whether maintaining green stems and upper leaves when water is limiting during
leaves under postanthesis drought increases grain yield in stay-green grain filling. Green leaf area at maturity is used as an
compared with senescent hybrids. Field studies were undertaken in indicator of postanthesis drought resistance in sorghum
northeastern Australia on a cracking and self-mulching gray clay. breeding programs in the USA (Rosenow et al., 1983)
Nine closely related hybrids varying in rate of leaf senescence were
and Australia (Henzell et al., 1992). Green leaf area at
grown under two water-limiting regimes, post-flowering water deficit
and terminal (pre- and postflowering) water deficit, and a fully irri-
maturity and its components have been found to vary
gated control. Under terminal water deficit, grain yield was correlated with both water regime and genotype (Borrell et al.,
positively with GLAM (r ⫽ 0.75**) and negatively with rate of leaf 2000). The critical issue is whether retention of green
senescence (r ⫽ ⫺0.74**). Grain yield also increased by ≈ 0.35 Mg leaf area under postanthesis drought actually increases
ha⫺1 for every day that onset of leaf senescence was delayed beyond grain yield in stay-green compared with senescent hy-
76 DAE in the water-limited treatments. Stay-green hybrids produced brids. Positive associations between green leaf area du-
47% more postanthesis biomass than their senescent counterparts ration and grain yield have been observed in a range
(920 vs. 624 g m⫺2) under the terminal water deficit regime. No differ- of cereals, including wheat, Triticum aestivum L. (Evans
ences in grain yield were found among eight of the nine hybrids under et al., 1975); maize, Zea mays L. (Tollenaar and Day-
fully irrigated conditions, suggesting that the stay-green trait did not
constrain yield in the well-watered control. The results indicate that
nard, 1978; Wolfe et al., 1988); oat, Avena sativa L.
sorghum hybrids possessing the stay-green trait have a significant
(Helsel and Frey, 1978); and sorghum (Henzell et al.,
yield advantage under postanthesis drought compared with hybrids 1992).
not possessing this trait. There is limited understanding of the physiological
processes underlying the stay-green trait, including the
basis of genetic variation. According to Bonhert et al.
(1995), mechanisms by which plants adapt to abiotic
W ater deficit is the major constraint to rainfed
sorghum production worldwide. Drought can oc-
cur before and after flowering, and resistance to water
stresses need to be quantified at a physiological, molecu-
lar, and genetic level, and future research must be di-
deficit at both of these stages has been reported in rected at functional characterization and biochemical
sorghum (Rosenow et al., 1996). Resistance to postan- integration of molecular and genetic data. Sorghum ge-
thesis drought is important in Australia’s northern grain notypes with the stay-green trait continue to fill their
belt, since crops generally grow into water deficit (Chap- grain normally under drought (Rosenow and Clark,
man et al., 2000). Symptoms of susceptibility to postan- 1981) and exhibit increased resistance to charcoal rot
thesis drought include premature leaf and stem senes- (Rosenow, 1984) and lodging (Henzell et al., 1984;
cence, charcoal rot [Macrophomina phaseolina (Tassi) Woodfin et al., 1988). Stay-green genotypes also contain
Goidanich], fusarium stalk rot (Fusarium moniliforme more cytokinins (McBee, 1984) and basal stem sugars
J. Sheld.), lodging, and reduced seed size. Expression (Duncan, 1984) than do senescent genotypes. Increased
of postanthesis drought symptoms is heightened when accumulation of soluble sugars in stay-green types is
crop growth is favorable prior to flowering and is fol- associated with greater functional leaf area during grain
lowed by severe water deficit, particularly in the latter filling, thereby reducing their dependence on stored as-
half of grain filling. Chapman et al. (2000) reported a similates from the stem to fill grain (Duncan et al., 1981,
McBee et al., 1983). Higher concentration of stem sugars
improves the digestible energy content of the straw,
A.K. Borrell and R.G. Henzell, Hermitage Research Station, Depart-
ment of Primary Industries, Warwick Queensland 4370, Australia;
G.L. Hammer, QDPI/CSIRO Agricultural Production Systems Re- Abbreviations: AGDM, aboveground dry mass; CGR, crop growth
search Unit, Toowoomba Queensland 4350, Australia. Received 24 rate; DAE, days after emergence; GLAM, green leaf area at maturity;
May 1999. *Corresponding author ([email protected]). ND, no water deficit treatment; PFD, postflowering water deficit
treatment; TD, terminal water deficit. *, **, *** Significant at the
Published in Crop Sci. 40:1037–1048 (2000). 0.05, 0.01, and 0.001 probability levels, respectively.
1038 CROP SCIENCE, VOL. 40, JULY–AUGUST 2000

making stay-green a valuable trait for both grain and soil type as Exp. 1. The experiment design was a randomized
fodder production in dual purpose sorghums (Van Oost- block design with four replicates and nine hybrids. Plot size
erom et al., 1996). If photosynthesis is maintained for was 6 by 1.42 m. The experiment site was fertilized prior to
longer than normal in stay-green types, they may yield sowing with 100 kg N ha⫺1 as urea. No irrigation was applied.
All plots were sown on 24 Nov. 1994 in rows 0.71 m apart
more in crops for which carbohydrate is a main harvest and subsequently thinned to a population of 100 000 plants
component (Thomas and Smart, 1993). ha⫺1. Totals of 123 and 148 mm of rain were recorded during
Initially, stay-green was selected under water-limited the pre- and postanthesis periods, with the largest fall (68 mm)
conditions to reduce lodging, since live plants have occurring 22 d after anthesis. A 6-m2 area (4.25 by 1.42 m)
stronger stems. However, previous research has found was machine harvested at maturity. Grain was dried in a forced
that such selection may reduce yield potential, as sor- draft oven at 80⬚C for 48 h before weighing.
ghum plants with a high grain sink/source ratio are more
likely to senesce when water is limiting (Henzell and Harvests
Gillieron, 1973; Rosenow et al., 1983). The need existed,
therefore, to clarify the association between rate of leaf In Exp. 1, a single row of length 1 m was cut from one of
the three center rows of each plot at 30, 46, 59 (anthesis ⫹
senescence and grain yield in water-limited environ-
3d), 87, and 114 d after emergence (DAE). At least 0.5-m
ments. To address this issue, Henzell et al. (1992) carried intervals of crop were left between sampling areas within a
out preliminary studies using visual ratings of leaf senes- row and no adjacent areas were sampled. Harvests at 30, 59,
cence. More detailed experiments on these associations and 114 DAE corresponded with the phenological stages of
are reported here. panicle differentiation, anthesis, and physiological maturity.
Our study had three main objectives. First, we exam- To determine the timing of panicle differentiation, two plants
ined grain yield and its components in nine closely re- per plot were sampled twice weekly from 21 DAE onwards.
lated hybrids varying in stay-green grown under three Preliminary stem dissection studies found that panicle differ-
water regimes. Second, we determined the association entiation, defined as the rapid elongation of the rachis and
the accompanying development of the upper primary and
between rate of leaf senescence and yield in these hy-
secondary branches (Moncur, 1981), corresponded with a pan-
brids under water-limiting conditions. Third, we exam- icle length of ≈ 2 mm. Hence panicle differentiation was de-
ined the partitioning of biomass among stem, leaf, and fined as the time when the panicle had attained a length of
panicle components in these hybrids under postanthesis 2 mm. Anthesis was defined as the time when 50% of the
water deficit, including the reliance of yield on stem anthers had extruded from 50% of 10 tagged panicles in each
reserves. Water and genotype effects on leaf area pro- plot. Physiological maturity was defined as the time at which
duction and senescence were examined in the first paper basal grains in 50% of the tagged panicles attained black layer.
of this series (Borrell et al., 2000). Each sample was dried in a forced draft oven at 80⬚C for 48 h
before weighing. All samples at 59 and 114 DAE, and also
TD samples at 87 DAE, were divided into mainstem and tiller
MATERIALS AND METHODS components, then further partitioned into green leaf, senesced
General leaf, stem (including leaf sheaths), and panicle. Plant number,
culm number, panicle number, aboveground dry mass
Details on the experiment site, treatments, agronomy and (AGDM), grain yield, 1000-grain weight, and stem length were
leaf observations are given in the first paper in this series determined at maturity (11 Apr. 1995) for all plots. Stem
(Borrell et al., 2000). Soil type is a cracking and self-mulching length was defined as the distance from the base of the stem
gray clay with abundant CaCO3 concretions and a high mont- to the top of the peduncle (characterized by the first branch
morillonite clay content (McKeown, 1978). Briefly, the experi- of the panicle).
ment design was a split plot with three replicates in which The following parameters were calculated on a plot basis.
three irrigation treatments were applied to main plots (6 by Harvest index was derived by dividing grain yield by
31.5 m) and nine hybrids varying in rate of leaf senescence aboveground dry mass. Grain number per panicle was calcu-
were allocated to subplots (3.5 by 6 m) (Exp. 1). All plots lated by dividing grain yield by the product of panicle number
were hand-sown under plastic covers on 15 Dec. 1994 and and mass per grain. Seasonal average crop growth rate was
emerged 18 Dec. 1994. The water regime treatments were no determined by dividing AGDM by the number of days from
deficit (ND), postflowering deficit (PFD), and terminal (pre- emergence to physiological maturity (black layer). Average
and postflowering) deficit (TD). No irrigation was applied to grain growth rate was calculated by dividing grain yield by
TD plots; plants in this treatment should have relied solely the number of days from anthesis to physiological maturity.
on stored soil water except that an additional 80 mm of water Stem reserves are defined as the difference in stem dry mass
entered the profile through the plastic in a series of rainfall between anthesis and maturity harvests.
events near anthesis. The magnitude of water entry under the
plastic was determined by the neutron scattering technique
Statistical Analyses
using a neutron probe (Model 503R, CPN Corp., Martinez,
CA) and there were no significant differences (P ⬎ 0.05) Data were analyzed by standard analysis of variance, and
among plots in water entry. The nine hybrids examined were pairwise comparisons of means were performed using the pro-
from crosses of three females varying in the B35 source of tected LSD procedure at P ⫽ 0.05 (Carmer and Swanson,
stay-green (AQL39, senescent; AQL41, intermediate; A35, 1973). Within each water treatment, correlations (n ⫽ 9) were
stay-green) and three males similarly varying in the KS19 calculated between grain yield and the following parameters:
source of stay-green (R69264, senescent; RQL36, intermedi- AGDM, harvest index, grain size, grain number per square
ate; RQL12, stay-green). meter, average crop growth rate, average grain growth rate,
The nine hybrids were also examined under rainfed condi- duration of grain growth, relative rate of leaf senescence (Bor-
tions at Hermitage Research Station (Exp. 2) on the same rell et al., 2000), GLAM, stem length, and stem reserves mobi-
 BORRELL ET AL.: STAY-GREEN AND DRY MATTER PRODUCTION IN SORGHUM 1039

lized for grain filling. Correlations were also calculated be-

tween green leaf dry mass at maturity and panicle dry mass
at maturity.

RESULTS AND DISCUSSION

Dry Matter Production
Genotype and water regime did not interact signifi-
cantly for biomass production at either panicle initiation
or anthesis. At panicle initiation (30 DAE), AGDM
was not affected by water regime (Fig. 1). However,
AGDM differed (P ⬍ 0.01) among genotypes, increas-
ing from 150 g m⫺2 (AQL39/RQL36) to 224 g m⫺2
(AQL39/R69264). By the anthesis harvest (59 DAE),
AGDM was similar in all genotypes, although variation
(P ⬍ 0.01) was observed among water regimes, with
AGDM being greater (P ⬍ 0.01) in ND and PFD (≈938
g m⫺2) compared with TD (828 g m⫺2).
A genotype ⫻ water regime interaction (P ⬍ 0.055)
was observed for biomass at maturity (Fig. 1). In the
five senescent hybrids (AQL39/R69264, AQL41/
R69264, A35/R69264, AQL39/RQL36, AQL41/RQL36;
open symbols), biomass production was almost 30% less
under TD compared with ND. Yet in the four stay-
green hybrids (A35/RQL36, AQL39/RQL12, AQL41/
RQL12, A35/RQL12; closed symbols), AGDM was only
13% less under TD compared with ND. For example,
when AQL39, AQL41, and A35 were crossed with
RQL36, AGDM in the senescent hybrid (AQL39/
RQL36) decreased from 2029 g m⫺2 (ND) to 1299 g
m⫺2 (TD), yet the decline was much less in the stay-
green hybrid (A35/RQL36), decreasing from 2204 g m⫺2
(ND) to 1827 g m⫺2 (TD).
Post-anthesis biomass production was 44% higher
(P ⬍ 0.001) for hybrids grown under ND compared with
TD (1089 vs. 756 g m⫺2). Genotypic variation was also
significant (P ⫽ 0.05), although the interaction between
genotype and water regime was not. Under terminal
water deficit, postanthesis biomass production in the
stay-green hybrid A35/RQL12 was twice that of the
senescent hybrid AQL39/R69264 (1029 vs. 537 g m⫺2).
Overall, stay-green hybrids produced 47% more bio-
mass between anthesis and maturity compared with
their senescent counterparts (920 vs. 624 g m⫺2) under
water-limited conditions (TD).

Rate of Crop Growth

There was no significant genotype ⫻ water regime
Fig. 1. Temporal pattern of aboveground dry matter production for
interaction for crop growth rate (CGR). Averaged nine sorghum hybrids grown under three water regimes: (a) no
across hybrids, CGR between emergence and maturity water deficit, (b) postflowering water deficit, and (c) terminal water
harvest (114 DAE) increased (P ⬍ 0.01) with water deficit. Anthesis at Day 56 is marked with an arrow. Vertical bars
regime: 14.4 (TD), 16.6 (PFD), and 17.9 (ND) g m⫺2 denote LSD (P ⫽ 0.05).
d⫺1. Averaged across water regimes, genotypic variation
in CGR for the same period was also significant (P ⬍ Partitioning of Biomass among Organs
0.01), ranging from 14.6 g m⫺2 d⫺1 (AQL39/RQL36) under Water-Limited Conditions
to 18.1 g m⫺2 d⫺1 (A35/RQL36). Since differences in The impact of rate of leaf senescence on biomass
phenology among genotypes were relatively small (Bor- partitioning among the leaf, stem, and panicle was exam-
rell et al., 2000), variation in biomass production was ined in the TD treatment, since severity of drought was
largely due to variation in CGR. greater in this treatment than in PFD. To demonstrate
1040 CROP SCIENCE, VOL. 40, JULY–AUGUST 2000

the effects of the B35 and KS19 sources of stay-green senescent (195 g m⫺2) and intermediate (232 g m⫺2)
on biomass partitioning, two examples will be discussed. hybrids (Fig. 2a). While green leaf dry mass declined
First, three females varying in rate of leaf senescence in all hybrids during late grain filling, the differences
(AQL39, senescent; AQL41, intermediate; A35, stay- observed at mid grain filling were maintained through
green) will be examined in crosses with a common male to maturity, resulting in more (P ⬍ 0.05) green leaf dry
(RQL36, intermediate) to highlight the impact of the mass at maturity in the stay-green hybrid (131 g m⫺2)
B35 source of stay-green on biomass partitioning. The than in the intermediate (58 g m⫺2) and senescent (72
effects of this source of stay-green on biomass parti- g m⫺2) hybrids. However, there was no difference in
tioning will also be examined in crosses with R69264 dead leaf dry mass among these genotypes (Fig. 2b).
(senescent) and RQL12 (stay-green) males. Second, Interestingly, stem dry mass remained relatively con-
three males similarly varying (R69264, senescent; stant (≈400 g m⫺2) throughout the grain-filling period in
RQL36, intermediate; RQL12, stay-green) will be ex- the stay-green hybrid, but declined in the intermediate
amined in crosses with a common female (AQL41, inter- hybrid from ≈400 to 300 g m⫺2 during the second half
mediate) to highlight the impact of the KS19 source of of the grain-filling period, such that stem dry mass was
stay-green on biomass partitioning. The effects of this significantly lower (P ⬍ 0.01) at maturity in the interme-
source of stay-green on biomass partitioning will also diate hybrid (Fig. 2c). Stem dry mass in the senescent
be examined in crosses with AQL39 (senescent) and hybrid was low (≈270 g m⫺2) throughout the grain-fill-
A35 (stay-green) females. ing period.
During the first half of the grain-filling period, the
Example 1 (B35 Source of Stay-Green)
accumulation of biomass in the panicle was similar in
During the first half of the grain-filling period, green the stay-green and intermediate hybrids and less in the
leaf dry mass remained relatively constant in AQL39/ senescent hybrid, although these differences were not
RQL36 (senescent) and AQL41/RQL36 (intermediate), significant (Fig. 2d). However, during the second half
yet increased in A35/RQL36 (stay-green) such that by of the grain-filling period, panicle dry mass increased
mid grain filling, green leaf dry mass in the stay-green by 204, 114, and 376 g m⫺2 in the senescent, intermediate,
hybrid (295 g m⫺2) was greater (P ⬍ 0.05) than in the and stay-green hybrids, respectively, resulting in almost

Fig. 2. The effects of the A35 source of stay-green on biomass partitioning among various plant components in crosses with RQL36: (a) green
leaf, (b) dead leaf, (c) stem, and (d) panicle. Vertical bars denote LSD (P ⫽ 0.05).
 BORRELL ET AL.: STAY-GREEN AND DRY MATTER PRODUCTION IN SORGHUM 1041

30% more (P ⬍ 0.05) panicle dry mass in the stay-green filling, green leaf dry mass declined faster in the senes-
than in the senescent hybrid. cent and intermediate hybrids compared with the stay-
There were no differences at maturity in green leaf green hybrid, resulting in greater (P ⬍ 0.05) green leaf
dry mass and panicle yield among the same three fe- dry mass at maturity in the stay-green hybrid (167 g
males (AQL39, AQL41, and A35) when crossed with m⫺2) than in the senescent and intermediate hybrids
R69264 (senescent male, data not shown). However, (≈56 g m⫺2). Conversely, dead leaf dry mass at maturity
when crossed with RQL12 (stay-green), green leaf dry was greater (P ⬍ 0.01) in the senescent and intermediate
mass and panicle dry mass at maturity were highly corre- hybrids (≈170 g m⫺2) than in the stay-green hybrid (89
lated (r ⫽ 0.86***, n ⫽ 9) in these females. Genotypic g m⫺2) (Fig. 3b).
differences in biomass partitioning were similar to those Throughout the grain filling period, stem dry mass
reported for crosses with RQL36, with green leaf dry remained relatively constant and high in the stay-green
mass at maturity and final panicle yield greater (P ⬍ hybrid (≈400 g m⫺2) and relatively constant and low
0.05) in the stay-green and intermediate hybrids com- in the senescent hybrid (≈320 g m⫺2) (Fig. 3c). The
pared with the senescent hybrid. Stem dry mass declined intermediate hybrid (AQL41/RQL36) is the same as
in the first half of the grain-filling period in the senescent that used in the first example and, as explained above,
hybrid, but remained consistently high in the stay-green its stem dry mass declined during the second half of the
and intermediate hybrids throughout the grain-filling grain-filling period. Green leaf dry mass at maturity was
period. correlated (r ⫽ 0.77**, n ⫽ 9) with final panicle dry
mass, resulting in greater grain yield in the stay-green
Example 2 (KS19 Source of Stay-Green) (1027 g m⫺2) than intermediate (756 g m⫺2) hybrid
During the first half of the grain-filling period, green (Fig. 3d).
leaf dry mass remained relatively constant in AQL41/ There were no differences at maturity in green leaf
R69264 (senescent) and AQL41/RQL36 (intermediate), dry mass and panicle yield among the same three males
but increased in AQL41/RQL12 (stay-green) so that by (R69264, RQL36, and RQL12) when crossed with
mid grain filling, green leaf dry mass in the stay-green AQL39 (senescent female, data not shown). However,
hybrid (263 g m⫺2) was greater (P ⬍ 0.05) than in the when crossed with A35 (stay-green), green leaf dry mass
senescent hybrid (198 g m⫺2) (Fig. 3a). During late grain and panicle dry mass at maturity were highly correlated

Fig. 3. The effects of the RQL12 source of stay-green on biomass partitioning among various plant components in crosses with AQL41: (a)
green leaf, (b) dead leaf, (c) stem, and (d) panicle. Vertical bars denote LSD (P ⫽ 0.05).
1042 CROP SCIENCE, VOL. 40, JULY–AUGUST 2000

(r ⫽ 0.96***, n ⫽ 9) in these males. Genotypic differ- hybrids, thereby enhancing stem strength and possibly
ences in biomass partitioning were similar to those re- reducing lodging.
ported for crosses with AQL41, with green leaf dry mass It has also been observed that stay-green genotypes
at maturity and final panicle yield greater (P ⬍ 0.05) produce significantly higher levels of sucrose, glucose,
in the stay-green and intermediate hybrids compared fructose, and starch within the plant, particularly in the
with the senescent hybrid. Stem dry mass declined stem, compared with senescent genotypes (McBee and
throughout the grain-filling period in the senescent hy- Miller, 1982; McBee et al., 1983; McBee, 1984). This
brid, remained consistently high in the intermediate hy- factor has been associated with increased yields in the
brid, and increased throughout the grain filling period stay-green genotypes, but the reason for this has not
in the stay-green hybrid, such that final stem dry masses been clear (McBee et al., 1983). The current study does
in the stay-green and intermediate hybrids were greater provide an explanation. Retention of the uppermost
(P ⬍ 0.05) than in the senescent hybrid (data not green leaves in stay-green hybrids during the latter half
shown). of the grain-filling period enables these types to con-
In both examples, most (⬎80%) of the increase in tinue assimilating C and complete grain filling, as evi-
panicle growth during the second half of the grain-filling denced by the positive correlation (r ⫽ 0.75**, n ⫽
period in the intermediate hybrid could be accounted 9) between GLAM and grain yield (Table 1). Various
for by reserves mobilized from the stem, assuming 100% workers (Stickler and Pauli, 1961; Goldsworthy, 1970;
conversion efficiency. However, since stem mass re- Fischer and Wilson, 1971) have shown the upper leaves
mained relatively constant during the grain-filling pe- contribute significantly to grain yield. In fact, Fischer
riod in the stay-green and senescent hybrids, it is likely and Wilson (1971) reported that assimilation by the
that panicle growth was largely dependent on photo panicle and upper four leaves accounted for 93% of
assimilation rather than stem reserves in these hybrids. grain yield. In our study, it is likely that stem masses
When crossed with RQL36, greater grain yield in the remained high in stay-green hybrids because sink de-
stay-green (A35/RQL36) than senescent (AQL39/ mand was largely met by current photo-assimilation,
RQL36) hybrid (933 vs. 669 g m⫺2) was probably associ- thereby minimizing the demand for nonstructural carbo-
ated with maintenance of photosynthetic capability in hydrate from the stem.
the stay-green hybrid, evidenced by more GLAM in the Is the under-utilization of stem and leaf reserves by
stay-green (17 459 cm2 m⫺2) than senescent (9575 cm2 stay-green hybrids a cost associated with this trait? As
m⫺2) hybrid. Similarly, when crossed with AQL41, the senescence is a mobilization function and a high harvest
greater grain yield in the stay-green hybrid (AQL41/ index is desirable in seed crops, domestication and vari-
RQL12) was associated with the retention of more green etal improvement have selected for efficient recovery
leaves during the latter half of grain filling. Overall, of nutrients from expendable, short-lifespan foliage
rates of crop growth during the latter half of the grain- (Thomas, 1992). Increased yield in stay-green hybrids
filling period in the stay-green hybrids (9.4 g m⫺2 d⫺1) is dependent on the supply of photo assimilate from
were twice that of the senescent hybrids (4.6 g m⫺2 d⫺1) green leaves exceeding the supply of preanthesis stem
under TD, providing evidence of continued photosyn- and leaf reserves in senescent and intermediate hybrids.
thetic activity in stay-green types. In the current studies, grain yield in stay-green hybrids
More direct evidence of extended photosynthetic ca- was greater than that in senescent hybrids by up to 25%
pability in stay-green compared with senescent sor- in Exp. 1 and 50% in Exp. 2, suggesting that there was
ghums is provided by De Villiers et al. (1993). Western no cost associated with nonsenescence.
analysis of proteins extracted from leaves found a high
degree of stability of chloroplast-associated proteins in Partitioning of Growth to Yield
two stay-green genotypes (Q101 and ICSV 745) com-
pared with a senescent genotype (R16). Q101, like Grain Yield
RQL12, is a QDPI line derived from KS19, which in Genotype and water regime interacted significantly
turn was derived from the cross between Combine Kafir (P ⬍ 0.055) for yield (Table 2). Grain yield declined with
60 and Short Kaura, the latter being from Nigeria. De- increasing water deficit in the five senescent hybrids
layed degradation of these proteins is correlated with (AQL39/R69264, AQL41/R69264, A35/R69264, AQL39/
the delayed onset of senescence in the stay-green geno- RQL36, AQL41/RQL36) but was maintained under in-
types. Indeed in Q101, the chloroplast proteins LHCP2, creasing water deficit in the four stay-green hybrids
OEC33, and Rubisco were retained until late in senes- (A35/RQL36, AQL39/RQL12, AQL41/RQL12, A35/
cence, indicating that photosynthesis may be maintained RQL12). For example, in AQL39/RQL36 (senescent),
for longer during senescence in this genotype. yield declined from 991 g m⫺2 under ND to 669 g m⫺2
Many earlier studies have reported an association under TD. In contrast, AQL41/RQL12 (stay-green)
between stay-green and lodging resistance (Henzell et maintained yield at ≈1000 g m⫺2 across all water regimes.
al., 1984; Rosenow, 1984; Woodfin et al., 1988). The Hence, grain yield under the fully irrigated control was
positive correlation (r ⫽ 0.71*, n ⫽ 9) reported in our not correlated with yield under terminal water deficit.
study between rate of leaf senescence and magnitude No variation in grain yield was observed among eight
of stem reserves mobilized under TD (Table 1) suggests of nine hybrids under ND (the exception being AQL41/
that stay-green hybrids are less reliant on nonstructural RQL36, which was greater than AQL39/R69264,
stem carbohydrate to fill their grain than are senescent AQL39/RQL36, AQL39/RQL12, AQL41/RQL12, and
 BORRELL ET AL.: STAY-GREEN AND DRY MATTER PRODUCTION IN SORGHUM 1043

Table 1. Correlation matrices for a range of yield determinants grown under three levels of water supply in Exp. 1 (n ⫽ 9).‡
Grain Grain Duration
Grain number growth of grain Rel. rate of Stem
Yield AGDM HI size m ⫺2 rate growth senescence GLAM reserves
No water deficit
Yield 1.00
AGDM‡ 0.90*** 1.00
HI 0.49 0.06 1.00
Grain size 0.06 ⫺0.23 0.61 1.00
Grain number/m2 0.77** 0.87*** 0.01 ⫺0.59† 1.00
Grain growth rate 0.95*** 0.95*** 0.26 ⫺0.08 0.81** 1.00
Duration of grain growth 0.29 ⫺0.04 0.73* 0.42 ⫺0.01 ⫺0.04 1.00
Rel. rate of senescence ⫺0.14 ⫺0.11 ⫺0.06 0.22 ⫺0.23 ⫺0.23 0.23 1.00
GLAM 0.57† 0.78** ⫺0.28 ⫺0.72* 0.92*** 0.66* ⫺0.20 ⫺0.31 1.00
Stem reserves ⫺0.31 ⫺0.58† 0.47 0.64* ⫺0.64* ⫺0.51 0.53† 0.62* ⫺0.79** 1.00
Postflowering water deficit
Yield 1.00
AGDM 0.74** 1.00
HI 0.64* 0.11 1.00
Grain size 0.42 0.21 0.82** 1.00
Grain number/m2 0.79** 0.69* 0.12 ⫺0.21 1.00
Grain growth rate 0.92*** 0.85*** 0.38 0.26 0.83** 1.00
Duration of grain growth 0.41 ⫺0.11 0.74** 0.46 0.11 0.02 1.00
Rel. rate of senescence 0.48 0.09 0.17 ⫺0.30 0.70* 0.38 0.36 1.00
GLAM ⫺0.30 0.25 ⫺0.50 ⫺0.22 ⫺0.15 ⫺0.13 ⫺0.47 ⫺0.63* 1.00
Stem reserves 0.10 ⫺0.28 0.62* 0.46 ⫺0.23 ⫺0.17 0.63* ⫺0.22 ⫺0.15 1.00
Terminal water deficit
Yield 1.00
AGDM 0.97*** 1.00
HI 0.39 0.15 1.00
Grain size 0.46 0.39 0.46 1.00
Grain number/m2 0.79** 0.80** 0.15 ⫺0.18 1.00
Grain growth rate 0.90*** 0.94*** 0.08 0.20 0.85*** 1.00
Duration of grain growth 0.70* 0.56 0.73* 0.70* 0.31 0.32 1.00
Rel. rate of senescence ⫺0.74** ⫺0.77** ⫺0.06 ⫺0.33 ⫺0.57† ⫺0.85*** ⫺0.20 1.00
GLAM 0.75** 0.82** ⫺0.09 0.34 0.57† 0.90*** 0.15 ⫺0.96*** 1.00
Stem reserves ⫺0.58† ⫺0.68* 0.26 0.18 ⫺0.76** ⫺0.74** ⫺0.02 0.71* ⫺0.71* 1.00
†, *, **, *** Significant at the 0.1, 0.05, 0.01, and 0.001 levels of probability, respectively.
‡ AGDM, aboveground dry matter; GLAM, green leaf area at maturity; HI, harvest index.

A35/RQL12), suggesting that in well-watered condi- under PFD (17 600 cm2 m⫺2) compared with ND (39 200
tions there was no yield cost associated with the stay- cm2 m⫺2). Relative rate of leaf senescence was not corre-
green trait. lated with yield, AGDM or harvest index (Table 1)
Grain yield under ND was highly correlated with because the onset of rapid senescence occurred too late
AGDM (r ⫽ 0.90***, n ⫽ 9), but not with harvest index, in the grain-filling period to greatly affect yield (Fig. 4).
indicating that production of biomass per se was more Rate of leaf senescence, however, when calculated from
important in yield attainment than partitioning of bio- the slope of the broken-stick function, was positively
mass to yield (Table 1). It is not surprising then, that correlated (r ⫽ 0.68*, n ⫽ 9) with grain yield. In this
grain yield under ND was also highly correlated with case, it is possible that grain yield determined the rate of
rate of crop growth (r ⫽ 0.82**, n ⫽ 9) and stem length
(r ⫽ 0.81**, n ⫽ 9) (data not shown). Of the components Table 2. Grain yield in nine sorghum hybrids grown under three
of yield, grain number per square meter was correlated levels of water supply in Exp. 1.
with grain yield (r ⫽ 0.77**, n ⫽ 9), but grain size Male parents
was not (Table 1). Furthermore, grain yield was not
Female parents R69264 RQL36 RQL12
correlated with any of the components of GLAM (total
plant leaf area, onset and rate of leaf senescence; data g m⫺2
not shown). Total plant leaf area was, however, nega- No water deficit
tively correlated (r ⫽ ⫺0.65*, n ⫽ 9) with onset of leaf AQL39 969 991 942
AQL41 1097 1193 969
senescence, indicating that leaves began to die earlier A35 1029 1071 1007
in those hybrids with a greater initial benchmark of Postflowering water deficit
green leaf area around anthesis. AQL39 996 792 986
Grain yield was not significantly less under PFD (1007 AQL41 1104 1073 991
g m⫺2) compared with ND (1030 g m⫺2). Since PFD A35 1081 1103 940
Terminal water deficit
plots were irrigated until just prior to anthesis, water
did not limit growth until late in the grain filling period. AQL39 753 669 827
AQL41 896 756 1027
Relative rate of leaf senescence was greater under PFD A35 844 933 936
compared with ND (1.13 vs. 0.32 percentage loss leaf LSD (0.05) ⫽ 174 (when comparing means within the same water regime)
LSD (0.05) ⫽ 171 (when comparing means among water regimes)
area index d⫺1), resulting in less than half the GLAM
1044 CROP SCIENCE, VOL. 40, JULY–AUGUST 2000

Table 3. Average grain yield of hybrids formed with three female

and three male parents varying in rate of senescence grown
under terminal water deficit (Exp. 1) and rainfed conditions
(Exp. 2).
Grain yield
(terminal deficit) Grain yield (rainfed)
g m⫺2
Female parents
AQL39 750 339
AQL41 893 549
A35 904 626
LSD (P ⫽ 0.05) 114 79
Male parents
R69264 831 488
RQL36 786 484
RQL12 930 543
LSD (P ⫽ 0.05) 114 NS

of leaf senescence per day (y ⫽ 1211 ⫺ 294x). In this

case, it is possible that rate of leaf senescence deter-
mined grain yield, and not vice versa, since onset of
Fig. 4. The relationship between onset of leaf senescence and grain
yield for sorghum hybrids grown under postflowering water deficit senescence commenced early enough for rate of senes-
and terminal water deficit. cence to affect yield.
The correlation between yield and onset of leaf senes-
senescence, and not vice versa, since high yield potential cence across the two water-limiting regimes (Fig. 4) has
had already been set before rapid senescence com- implications for plant breeders who visually rate stay-
menced (Fig. 4), and leaves senesced in response to green at maturity only. For example, a subset of hybrids
sink demand. may receive the same stay-green rating at maturity, yet
Yields were still relatively high within the TD treat- large variation in yield may also be observed within this
ment, primarily due to infiltration of ≈80 mm of water subset due to differences in the onset of leaf senescence
under the plastic during a large rainfall event near anthe- among hybrids. Therefore some means of rapidly as-
sis. Despite this water entry, plants grown under TD sessing the onset of leaf senescence may be useful in
encountered severe water stress during the second half screening hybrids for stay-green. Ultimately, selection
of the grain-filling period, resulting in yields ranging for the stay-green phenotype based on molecular mark-
from 669 g m⫺2 (AQL39/RQL36, senescent) to 1027 g ers associated with delayed onset of leaf senescence
m⫺2 (AQL41/RQL12, stay-green). That hybrids grown should be worthwhile.
under PFD and TD displayed similar leaf senescence Although differential expression of stay-green is gen-
patterns (Borrell et al., 2000) despite PFD hybrids yield- erally observed in sorghum crops yielding ⬍4 Mg ha⫺1
ing more than TD hybrids (Table 2) appears anomalous. under postanthesis drought, this trait can still confer a
A closer examination of the senesced plant leaf area significant advantage at higher yield levels. The critical
functions for PFD and TD (Borrell et al., 2000) reveals issue is the timing and severity of drought in relation
that onset of senescence was delayed (P ⬍ 0.1) and rate to crop growth and water supply, rather than yield po-
of senescence was greater (P ⬍ 0.001) in hybrids grown tential per se. Yield under TD was greater (P ⬍ 0.05)
under PFD compared with TD. Of these components in A35 hybrids (904 g m⫺2) and AQL41 hybrids (893 g
of senescence, onset was more important than rate in m⫺2) than in AQL39 hybrids (750 g m⫺2) (Table 3).
explaining yield variation across these treatments. A Yield was also greater (P ⬍ 0.05) in RQL12 hybrids
combined analysis of PFD and TD data found that grain (930 g m⫺2) than in R69264 hybrids (831 g m⫺2) and
RQL36 hybrids (786 g m⫺2) under TD conditions.
yield was correlated with onset of leaf senescence (r ⫽
In the complementary rainfed study (Exp. 2), grain
0.653**, n ⫽ 18, Fig. 4), but not with rate. This relation-
yields were much lower (3–6 Mg ha⫺1, Tables 3 and 4),
ship shows that grain yield increased by ≈0.35 Mg ha⫺1
but yield trends among genotypes were similar for plants
for every day (or 3.1 g m⫺2 ⬚C d⫺1) that onset of senes-
grown under TD in Exp. 1. Plots were irrigated prior
cence was delayed beyond 76 DAE (866 ⬚C d). The
to sowing in Exp. 1, yet Exp. 2 was not irrigated and
extent to which delayed onset of leaf senescence affects
grain yield will depend on the timing and severity of Table 4. Grain yield of nine sorghum hybrids grown under rainfed
drought. Hence in Exp. 1, the greater yield in PFD than conditions (Exp. 2).
TD hybrids (1007 vs. 849 g m⫺2) is explained largely by
Male parents
the delayed onset of senescence in PFD compared with Female
TD hybrids (943 vs. 914 ⬚C d). Within TD, grain yield parents R69264 RQL36 RQL12 Mean
was correlated with onset of leaf senescence (r ⫽ 0.615*, g m⫺2
n ⫽ 9), but not with rate (r ⫽ ⫺0.345, n ⫽ 9). Yield AQL39 303 323 390 339
AQL41 567 504 577 549
was negatively correlated with relative rate of leaf senes- A35 594 623 661 626
cence (r ⫽ ⫺0.74**, n ⫽ 9), equating to a yield decline Mean 488 483 543
LSD (0.05) ⫽ 137 LSD (0.01) ⫽ 185
of ≈30 g m⫺2 for each 0.1% increase in the relative rate
 BORRELL ET AL.: STAY-GREEN AND DRY MATTER PRODUCTION IN SORGHUM 1045

supports the above hypothesis of Thomas and Smart

(1993), since grain yield variation among hybrids under
TD was correlated with green leaf area at maturity (r ⫽
0.75**) and relative rate of leaf senescence (r ⫽
⫺0.74**) (Table 1). These outcomes agree with earlier
work by Henzell et al. (1992) which found a significant
negative correlation between visual leaf senescence rat-
ings and yield in 76 grain sorghum hybrids grown under
water-limited conditions in central Queensland.

Harvest Index
There was no genotype ⫻ water regime interaction
for harvest index. Harvest index was greater (P ⬍ 0.05)
for plants grown under PFD (0.54) and TD (0.53) com-
pared with those grown under ND (0.51). Genotypic
variation in harvest index was highly significant (P ⬍
0.01), ranging from 0.50 (AQL39/RQL36 and A35/
RQL36) to 0.55 (AQL41/R69264).
Under TD, the relative rate of leaf senescence was
correlated with AGDM (r ⫽ ⫺0.77***), but not with
Fig. 5. The relationship between relative rate of leaf senescence and
stem reserves mobilized for nine sorghum hybrids grown under harvest index (Table 1), suggesting the association be-
terminal water deficit. tween high grain sink/source ratio and senescence under
water-limited conditions reported by Henzell and Gil-
grain yield relied largely on in-crop rainfall. Further- lieron (1973), Duncan et al. (1981), Rosenow et al.
more, the heavy rainfall in February coincided with an- (1983), and Tangpremsri (1989) can be broken. Indeed,
thesis in Exp. 1 and with mid grain filling in Exp. 2, AQL41/RQL12 attained a high grain yield under termi-
thereby substantially enhancing grain growth in Exp. 1, nal deficit by combining a low rate of leaf senescence
yet having less impact on yield in Exp. 2. Grain yield with a high harvest index. This outcome is encouraging,
ranged (P ⬍ 0.01) from 303 g m⫺2 (AQL39/R69264, and allays concerns expressed by some plant breeders
highly senescent) to 661 g m⫺2 (A35/RQL12, highly stay- that selection for stay-green may result in developing
green). Yield of A35 hybrids (626 g m⫺2) was greater hybrids with small panicles. The observation that in
than for AQL39 hybrids (339 g m⫺2) and AQL41 hybrids dioecious plants such as Spinacia and Cannabis, the
(549 g m⫺2). No differences in grain yield were observed leaves of the two sexes senesce simultaneously during
among the male parents (Table 3). the reproductive phase despite the difference in sink
Thomas and Smart (1993) suggested that plants exhib- load between seedless males and fruit-bearing females,
iting Type A stay-green (delayed onset of senescence) provides further evidence that retention of leaf green-
and Type B stay-green (reduced rate of senescence) ness is not solely dependent on sink demand
might be expected to show a higher yield in crops for (Thomas, 1992).
which carbohydrate is a major component of the harvest,
since these stay-green types continue to photosynthesise Contribution of Stem Reserves to Grain Yield
for longer than normal. Borrell et al. (2000) found the The contribution of stem reserves (defined as the
KS19 source of stay-green displays both Types A and difference in stem dry mass between anthesis and matu-
B, although the B35 source of stay-green displays only rity harvests) to grain yield was affected by water regime
Type B behavior; that is, the onset of leaf senescence and hybrid. There are a number of weaknesses inherent
is delayed compared with the normal type although to the method of determining C reserves by dry weight
the rate of leaf senescence is not different. Our study analysis (Borrell et al., 1989). First, no allowance is made

Table 5. Grain yield, grain number per square meter, mass per grain, and duration and rate of grain growth for nine sorghum hybrids
averaged across three water regimes in Exp. 1.
Grain Mass per Duration of Rate of
Hybrid Grain yield number grain grain growth grain growth
g m⫺2 m ⫺2 mg d g m⫺2 d⫺1
AQL39/R69264 906 31 331 29.1 55.8 16.3
AQL41/R69264 1032 38 232 27.0 59.2 17.5
A35/R69264 985 32 986 30.0 59.6 16.5
AQL39/RQL36 817 33 538 24.4 54.2 15.0
AQL41/RQL36 1007 38 070 26.4 55.9 17.9
A35/RQL36 1035 38 816 26.8 55.8 18.6
AQL39/RQL12 918 36 155 25.4 54.1 17.0
AQL41/RQL12 996 34 428 28.9 57.7 17.3
A35/RQL12 961 34 446 27.9 55.9 17.2
LSD (P ⫽ 0.05) 100 3 899 1.4 2.5 1.9
1046 CROP SCIENCE, VOL. 40, JULY–AUGUST 2000

for any decline in stem mass due to respiration; second, m⫺2 (A35/R69264). Under PFD, relative rate of leaf
there is no estimate of the allocation of the dry matter senescence (percentage loss leaf area index d⫺1, Borrell
transfer to any other organs; third, maximum stem mass et al., 2000) was not correlated with the magnitude of
may be attained after anthesis; and fourth, it is possible stem reserves mobilized (Table 1).
to vary the apparent contribution of stem reserves to However, under TD all hybrids except A35/RQL12
grain by varying grain yield. Nonetheless, dry matter mobilized some stem reserves during grain filling. In
analysis of the stem and grain does provide an estimate absolute terms, the estimated contribution of stem re-
of the contribution of stem reserves to grain yield. serves to grain yield under TD was small, varying from
Between anthesis and maturity, stem dry mass in- zero in A35/RQL12 to about 100 g m⫺2 (15% of grain
creased (P ⬍ 0.05) by an average of 36 g m⫺2 across all yield) in AQL39/R69264 and AQL41/RQL36. Under
hybrids under fully irrigated conditions, and therefore TD, relative rate of leaf senescence was positively corre-
stem reserves did not contribute to yield under ND. lated (r ⫽ 0.71*, n ⫽ 9) with the magnitude of stem
Under PFD, biomass accumulation in the stem during reserves mobilized (Fig. 5), indicating that stay-green
the grain-filling period varied among hybrids, increasing hybrids (filled symbols) do not deplete stem reserves
in four and decreasing in five hybrids. For the latter to fill grain to the same extent as senescent hybrids
hybrids, the estimated contribution of stem reserves to (open symbols), and suggests that they may be more
grain yield ranged from 7 g m⫺2 (A35/RQL12) to 22 g resistant to lodging.

Components of Yield
There was no significant genotype ⫻ water regime
interaction for grain number per square meter. Grain
numbers were lower (P ⬍ 0.01) under TD (30 720) com-
pared with ND and PFD (37640), and ranged among
genotypes from 31 330 (AQL39/R69264) to 38 820 (A35/
RQL36) (Table 5). Genotype and water regime inter-
acted (P ⬍ 0.1) for panicle number per square meter
(data not shown). For example, the number of panicles
declined from 21 m⫺2 (ND) to 14 m⫺2 (TD) in AQL39/
RQL36 (senescent), yet remained at ≈19 m⫺2 across all
water regimes in A35/RQL36 (stay-green), suggesting
that tiller survival may be enhanced by stay-green in
some genetic backgrounds. No genotype ⫻ water regime
interaction was observed for grain number per panicle,
with a value of ≈2000 across all water regimes (data not
shown). However, genotypes varied (P ⬍ 0.01) in this
parameter, ranging from 1660 (AQL39/R69264) to 2246
(A35/RQL12).
No significant genotype ⫻ water regime interaction
was observed for grain size. Averaged across hybrids,
grain size was unaffected by water regime, maintaining
a value of ≈27 mg across all treatments. Averaged across
water regimes, genotypic variation was significant (P ⬍
0.01), ranging from 24.4 mg (AQL39/RQL36) to 30.0
mg (A35/R69264) (Table 5). Overall, grain yield was
correlated (P ⬍ 0.01) with grain number per square
meter in all water regimes, but was not correlated with
grain size in any water regime (Table 1). Furthermore,
grain growth rates were simply a function of grain num-
bers, since grain growth rate was correlated with grain
number per square meter in ND (r ⫽ 0.81**, n ⫽ 9),
PFD (r ⫽ 0.83**, n ⫽ 9) and TD (r ⫽ 0.85***, n ⫽ 9), but
was not correlated with grain size in any water regime.

Duration and Rate of Grain Growth

Genotype and water regime did not significantly in-
teract for duration of grain growth (anthesis to physio-
logical maturity). The length of the grain-filling period
Fig. 6. The relationships between (a) relative rate of leaf senescence was ≈56 d for all water regimes. However, considerable
and grain growth rate and (b) grain growth rate and grain yield genotypic variation was observed (P ⬍ 0.01), ranging
for nine sorghum hybrids grown under terminal water deficit. from 54 d (AQL39/RQL36 and AQL39/RQL12) to 60 d
 BORRELL ET AL.: STAY-GREEN AND DRY MATTER PRODUCTION IN SORGHUM 1047

(A35/R69264) (Table 5). Short grain-filling periods in increased yield and lodging resistance, without imposing
AQL39/RQL36 and AQL39/RQL12 were primarily as- a yield cost in wetter years.
sociated with late flowering and early maturity, respec-
tively (Borrell et al., 2000). ACKNOWLEDGMENTS
There was no significant genotype ⫻ water regime
Andrew Douglas is thanked for his technical assistance
interaction for rate of grain growth. Average rate of during this study, and Kerry Bell and David Butler for their
grain growth was higher (P ⬍ 0.01) under ND (18.1 g assistance in statistical analysis. The contribution to this re-
m⫺2 d⫺1) compared with TD (15.4 g m⫺2 d⫺1). Genotypic search by farm staff at Hermitage Research Station is grate-
variation was also observed, ranging (P ⬍ 0.05) from fully acknowledged. This work was funded by the Farming
15.0 g m⫺2 d⫺1 (AQL39/RQL36) to 18.6 g m⫺2 d⫺1 (A35/ Systems Institute of the Queensland Department of Primary
RQL36) (Table 5). Industries and the Grains Research and Development Corpo-
Since the stay-green trait extends green leaf area du- ration.
ration in the latter half of grain filling, it might be ex-
pected that the length of the grain-filling period be in- REFERENCES
creased accordingly, but the duration of grain filling was Bonhert, H.J., D.E. Nelson, and R.G. Jensen. 1995. Adaptations to
not correlated with relative rate of leaf senescence under environmental stresses. Plant Cell 7:1099–1111.
TD. Rather, relative rate of leaf senescence was corre- Borrell, A.K., G.L. Hammer, and A.C.L. Douglas. 2000. Does main-
taining green leaf area in sorghum improve yield under drought?
lated with grain growth rate (r ⫽ ⫺0.85***) and, in turn, I. Leaf growth and senescence. Crop Sci. 40:1026–1037 (this issue).
grain growth rate was correlated with grain yield (r ⫽ Borrell, A.K., L.D. Incoll, R.J. Simpson, and M.J. Dalling. 1989. Parti-
0.90***, Fig. 6). Therefore rate, and not duration of tioning of dry matter and the deposition and use of stem reserves
grain growth, was the more important factor in yield in a semi-dwarf wheat crop. Ann. Bot. 63:527–539.
Carmer, S.G., and M.R. Swanson. 1973. An evaluation of ten pairwise
determination under TD. multiple comparison procedures by Monte Carlo methods. J. Amer.
Stat. Assoc. 68:66–74.
Chapman, S.C., M. Cooper, G.L. Hammer, and D. Butler. 2000. Geno-
CONCLUSIONS type by environment interactions affecting grain sorghum. II. Fre-
quencies of different seasonal patterns of drought stress are related
Postanthesis biomass production was 44% higher for to location effects on hybrid yields. Aust. J. Agric. Res. 51:209–222.
hybrids grown under fully irrigated conditions com- De Villiers, L., K. Turk, H. Thomas, and C. Howarth. 1993. Analysis
pared with terminal water deficit (1089 vs. 756 g m⫺2). and exploitation of the staygreen character in sorghum. Overseas
Under terminal water deficit, grain yield was correlated Development Administration Project R4885, Annual Report 1993.
Duncan, R.R. 1984. The association of plant senescence with root
positively with GLAM and negatively with rate of leaf and stalk disease in sorghum. p. 99–100. In L.K. Mughogho (ed.)
senescence. Importantly, harvest index was not corre- Sorghum root and stalk diseases, a critical review. Proc. Consulta-
lated with rate of leaf senescence, allaying concerns tive group discussion of research needs and strategies for control
expressed by some plant breeders that selection for stay- of sorghum root and stalk diseases. Bellagio, Italy. 27 Nov.–2 Dec.
1983. ICRISAT, Patancheru, A.P., India.
green may result in developing hybrids with small pani- Duncan, R.R., A.J. Bockholt, and F.R. Miller. 1981. Descriptive com-
cles. Grain yield also increased by ≈0.35 Mg ha⫺1 for parison of senescent and nonsenescent sorghum genotypes. Agron.
every day that onset of leaf senescence was delayed J. 73:849–853.
beyond 76 DAE under water-limited conditions. Stay- Evans, L.T., I.F. Wardlaw, and R.A. Fischer. 1975. Wheat. p. 101–150.
In L.T. Evans (ed.) Crop physiology: Some case histories. Cam-
green hybrids produced 47% more biomass between bridge University Press, Cambridge, UK.
anthesis and maturity than their senescent counterparts Fischer, K.S., and G.L. Wilson. 1971. Studies of grain production in
(920 vs. 624 g m⫺2) under water-limited conditions, sug- Sorghum vulgare. 1. The contribution of pre-flowering photosyn-
gesting that sorghum hybrids possessing the stay-green thesis to grain yield. 2. Sites responsible for grain dry matter produc-
tion during the post-anthesis period. Aust. J. Agric. Res. 22:33–47.
trait have a significant yield advantage under postanthe- Goldsworthy, P.R. 1970. The sources of assimilate for grain develop-
sis drought. ment in tall and short sorghum. J. Agric. Sci. 74:523–531.
No variation in grain yield was observed among eight Helsel, D.B. and K.J. Frey. 1978. Grain yield variations in oats associ-
of nine hybrids in the fully irrigated control, suggesting ated with differences in leaf area duration among oat lines. Crop
Sci. 18:765–769.
that there was no yield cost associated with the stay- Henzell, R.G., R.L. Brengman, D.S. Fletcher, and A.N. McCosker.
green trait under well-watered conditions. Grain yields 1992. Relationships between yield and non-senescence (stay-green)
under ND and TD were not significantly correlated. in some grain sorghum hybrids grown under terminal drought
This study also provides evidence that stay-green may stress, p. 355–358. In M.A. Foale, R.G. Henzell and P.N. Vance
(ed.) Proceedings of the Second Australian Sorghum Conference,
contribute to lodging resistance. A positive correlation Gatton, Feb 4–6 1992. Australian Institute of Agricultural Science,
was observed between rate of leaf senescence and mag- Melbourne. Occasional Publication No. 68.
nitude of stem reserves mobilized, suggesting that stay- Henzell, R.G., R.L. Dodman, A.A. Done, R.L. Brengman, and P.E.
green hybrids are less reliant on nonstructural stem car- Mayers. 1984. Lodging, stalk rot, and root rot in sorghum in Austra-
lia, p. 225–235. In L.K. Mughogho (ed.) Sorghum root and stalk
bohydrate to fill their grain than are senescent hybrids. diseases, a critical review. Proc. Consultative group discussion of
This may result in stronger stems and less lodging. research needs and strategies for control of sorghum root and
Overall, stay-green hybrids yielded more than senes- stalk diseases. Bellagio, Italy. 27 Nov.–2 Dec. 1983. ICRISAT,
cent hybrids under postanthesis drought, yet there was Patancheru, A.P., India.
Henzell, R.G., and W. Gillieron. 1973. Effects of partial and complete
no yield cost associated with this trait under well-wa- panicle removal on the rate of death of some Sorghum bicolor
tered conditions. Therefore, growing stay-green hybrids genotypes under moisture stress. Qld. J. Agric. Animal. Sci. 30:
should benefit sorghum producers in drier years through 291–299.
1048 CROP SCIENCE, VOL. 40, JULY–AUGUST 2000

ICRISAT. 1988. Environmental challenges: some solutions. ICRISAT p. 18–31. In H.D. Loden and D. Wilkinson (ed.) Proc. 36th Annu.
Research Highlights 1988, p. 9–10. International Crops Research Corn and Sorghum Industry Res. Conf. Chicago, IL. 9–11 Dec.
Institute for the Semi-Arid Tropics, Patancheru, A.P. 502 324, 1981. Am. Seed Trade Assoc., Washington, DC.
India. Rosenow, D.T., J.W. Johnson, R.A. Frederiksen, and F.R. Miller.
Katsanos, R.A., and A.J. Pappelis. 1965. Seasonal trends in density 1977. Relationship of nonsenescence to lodging and charcoal rot
and cell death in sorghum stalk tissue. Phytopathology 55:97–99. in sorghum. p. 69. In Agronomy Abstracts. ASA, Madison, WI.
McBee, G.G. 1984. Relation of senescence, nonsenescence, and kernel Rosenow, D.T., J.E. Quisenberry, C.W. Wendt, and L.E. Clark. 1983.
maturity to carbohydrate metabolism in sorghum, p. 119–129. In Drought tolerant sorghum and cotton germplasm. Agric. Water
L.K. Mughogho (ed.) Sorghum root and stalk diseases, a critical Manage. 7:207–222.
review. Proc. Consultative group discussion of research needs and Rosenow, D.T., G. Ejeta, L.E. Clark, M.L. Gilbert, R.G. Henzell,
strategies for control of sorghum root and stalk diseases. Bellagio, A.K. Borrell, and R.C. Muchow. 1996. Breeding for pre- and post-
Italy. 27 Nov.–2 Dec. 1983. ICRISAT, Patancheru, A.P., India. flowering drought stress resistance in sorghum. p. 400–411. In D.T.
McBee, G.G., and F.R. Miller. 1982. Carbohydrates in sorghum culms Rosenow and J.M. Yohe (ed.) Proceedings of the International
as influenced by cultivars, spacing, and maturity over a diurnal Conference on Genetic Improvement of Sorghum and Pearl Millet.
period. Crop Sci. 22:381–385. Lubbock, TX. 22–27 Sept. 1996. INTSORMIL, Lubbock, TX.
McBee, G.G., R.M. Waskom III, F.R. Miller, and R.A. Creelman. ICRISAT, India.
1983. Effect of senescence and nonsenescence on carbohydrates Stickler, F.C., and A.W. Pauli. 1961. Leaf removal in grain sorghum.
in sorghum during late kernel maturity states. Crop Sci. 23:372–376. 1. Effect of certain defoliation treatments on yield and components
McKeown, F.R. 1978. A land classification of the Hermitage Research of yield. Agon. J. 53:99–102.
Station. Division of Land Utilisation. Queensland Department of Tangpremsri, T. 1989. Physiological studies on the performance of
Primary Industries. sorghum genotypes under moisture limiting conditions. Ph.D. the-
Moncur, M.W. Grain sorghum. p. 24–25. In Floral initiation in field sis. Univ. of Queensland, Brisbane, Australia.
crops: An atlas of scanning electron micrographs. Commonwealth Thomas, H. 1992. Canopy survival. p. 11–41. In N.R. Baker and H.
Scientific and Industrial Research Organisation, Melbourne, Aus- Thomas (ed.) Crop photosynthesis: Spatial and temporal determi-
tralia. nants. Elsevier, Amsterdam.
Passioura, J.B. 1986. Resistance to drought and salinity: Avenues for Thomas, H., and C.M. Smart. 1993. Crops that stay green. Ann. Appl.
improvement. Aust. J. Plant Physiol. 13:191–201. Biol. 123:193–219.
Rosenow, D.T. 1977. Breeding for lodging resistance in sorghum. p. Tollenaar, M., and T.B. Daynard. 1978. Leaf senescence in short-
171–185. In H.D. Loden and D. Wilkinson (ed.) Proceedings of season maize hybrids. Can. J. Plant Sci. 58:869–874.
the 32nd Annual Corn and Sorghum Research Conference. Am. Van Oosterom, E.J., R. Jayachandran, and F.R. Bidinger. 1996. Diallel
Seed Trade Assoc. Washington, DC. analysis of the stay-green trait and its components in sorghum.
Rosenow, D.T. 1984. Breeding for resistance to root and stalk rots Crop Sci. 36:549–555.
in Texas. p. 209–217. In L.K. Mughogho (ed.) Sorghum root and Wolfe, D.W., D.W. Henderson, T.C. Hsiao, and A. Alvino. 1988.
stalk diseases, a critical review. Proc. Consultative group discussion Interactive water and nitrogen effects on senescence of maize. I.
of research needs and strategies for control of sorghum root and Leaf area duration. Agron. J. 80:859–864.
stalk diseases. Bellagio, Italy. 27 Nov.–2 Dec. 1983. ICRISAT, Woodfin, C.A., D.T. Rosenow, and L.E. Clark. 1988. Association
Patancheru, A.P., India. between the stay-green trait and lodging resistance in sorghum. p.
Rosenow, D.T., and L.E. Clark. 1981. Drought tolerance in sorghum. 102. In Agronomy Abstracts. ASA, Madison, WI.