Tải bản đầy đủ - 0trang
II. Seed Filling: Definition and Measurement
D. B. EGLI
Both of these methods can be applied at all levels, but the growth curve
method requires extensive sampling and selection of the appropriate regression
model to produce acceptable estimates. The EFP avoids dealing with the nonlinear lag phases of seed growth and requires fewer samples.
Plant growth stages can also be used to estimate seed-ﬁll duration at the
individual plant or plant community levels. The time between growth stages R5
(beginning of seed ﬁll) and R7 (an estimate of PM, Fehr and Caviness, 1977) is
widely used in soybean [Glycine max (L.) Merr.] (Egli et al., 1984; Agudelo et al.,
1986). The period between anthesis or silking and PM is popular in many cereals
(Daynard and Kannenberg, 1976; Darrock and Baker, 1995). PM is usually
estimated from a visual appraisal of seed or plant characteristics (Egli, 1998).
Growth stage estimates may not be as precise as those based on seed growth
curves [e.g., R5 – R7 is not the same for indeterminate and determinate soybean
cultivars (Agudelo et al., 1986; Pfeiffer and Egli, 1988)], but they have the
advantage in that they are non-destructive and easy to determine.
Munier-Jolain et al. (1993) and Ney et al. (1993) used seed moisture levels,
which are closely associated with the stage of seed development (Fraser et al.,
1982), to estimate seed-ﬁll duration in soybean and pea (Pisum sativum L.).
Comparisons of published estimates of seed-ﬁll duration can be misleading
because of the variety of methods used to estimate it, species variation in growth
characteristics and environmental effects. However, comparisons of published
estimates of seed-ﬁll duration provide some useful general information about the
length and variation of this important growth stage. The variation in seed-ﬁll
duration among species, based on whole-plant growth characteristics, is
frequently less than the variation within a species in a summary of ﬁeld trials
(Table I). A similar summary of EFP also found more variation within than
between species (Egli, 1981). The longer seed-ﬁll duration of maize (Zea mays
L.) (Table I) represents an apparent exception to this pattern, however, this
advantage was not evident in comparisons of EFP (Egli, 1981), so it could be
longer simply because the anthesis to beginning seed growth period is longer in
maize than in other species. There are no obvious relationships between known
plant characteristics and seed-ﬁll duration in Table I. The seed-ﬁll duration of C4
crops [sorghum (Sorghum bicolor L.), pearl millet (Pennisetum glaucum (L.)
Leeke)]—discounting any advantage for maize—is not longer than C3 crops.
Species with high seed protein [soybean, groundnut (Arachis hypogea L.)] or oil
[sunﬂower (Helianthus annuus L.), groundnut] concentrations do not have
shorter seed-ﬁll durations than cereals [wheat (Triticum aestivum L.), rice (Oryza
sativa L.)] that produce seeds with low protein and oil concentrations.
The seed-ﬁlling period represents less than half of the total growth cycle of
most crops. It represented only 26 –41% of the total growth cycle of a group
of soybean cultivars from Maturity Groups 00 through V (Zeiher et al., 1982;
Egli, 1994). The proportion was only 32 – 38% in pearl millet (Craufurd and
Bidinger, 1988), but it was slightly larger in maize (44% for six hybrids)
Variation in Seed-Fill Duration Within and Among Crop Species
Silking to PM
Silking to PM
Silking to maximum seed mass
Anthesis to maturity
Anthesis to PM
Broken stick-iterative regression
Flowering to PM
Flowering to maturity
First anthesis to PM
Growth stage R5–R7
Growth stage R5–R7
Sigmoid growth curve
Flowering to maturity
Daynard and Kannenberg (1976)
Hanway and Russell (1969)
Gebeyehou et al. (1982)
Reynolds et al. (1994)
Garcia del Moral et al. (1991)
Jones et al. (1979)
Craufurd and Bidinger (1988)
Villalobos et al. (1994)
Boerma and Ashley (1988)
Davies et al. (1999)
Dow el-madina and Hall (1986)
Witzenberger et al. (1988)
PM, physiological maturity.
Approximate concentrations from Sinclair and de Wit (1975), Hulse et al. (1980), Langer and Hill (1991) and Bewley and Black (1994).
Range based on cultivar means.
Diverse genotypes from four countries grown in California.
D. B. EGLI
(Hanway and Russell, 1969). The preliminary events in the yield production
process (vegetative growth, ﬂowering and seed set), take up to twice as much
time as the production of the seeds that make up economic yield. All crops do not
suffer from such a limited yield production period—tuber growth in potato
(Solanum tuberosum spp. tuberosum) continues for up to 120 days (Moorby and
Milthorpe, 1975; Spitters, 1987; Vos and Groenwold, 1987; Burton, 1989).
Storage root growth in sugar beet (Beta vulgaris L.) covers 90– 110 days (Milford
and Watson, 1971; Fick et al., 1975) and cassava (Manihot esculenta Crantz)
roots accumulate dry matter for up to 300 days (Howeler and Cadavid, 1983;
Aleves, 2002). Compared to these root and tuber crops, the time allocated to yield
production in grain crops is not long, which puts much more emphasis on the rate
of growth when producing exceptionally high yields.
III. VARIATION IN SEED-FILL DURATION
Seed-ﬁll duration, like most plant growth processes, has both an environmental and a genetic component. The length of the seed-ﬁlling period of a crop in
a speciﬁc environment is determined by the genetic potential of the genotype and
the environment. Environmental conditions during vegetative growth, ﬂowering
or seed set could indirectly affect seed-ﬁll duration by altering other aspects
of plant growth, but many environmental effects are a direct response to the
environment during seed ﬁlling. We will focus on direct effects of the
environment during seed ﬁlling.
Seed-ﬁll duration generally increases as temperatures decrease below 308C in
many crops (Fig. 1) including maize (Tollenaar and Bruulsema, 1988; Muchow,
1990; Wilhem et al., 1999), spring and winter wheat (Spiertz, 1978; Vos, 1981;
Al-Khatib and Paulsen, 1984; Wardlaw and Moncur, 1995; Gibson and Paulsen,
1999), sunﬂower (Chimenti et al., 2001), lentil (Lens culinais) (Summerﬁeld
et al., 1989) and oat (Avena sativa L.) (Hellewell et al., 1996). There may be
exceptions to this general trend, for example, there was little effect of
temperatures between 20 and 308C on seed-ﬁll duration in soybean (Egli and
Wardlaw, 1980) or rice (Chowdhury and Wardlaw, 1978). The seed-ﬁll duration
of rice and wheat was the same at 36/318C (day/night temperature), but the
duration of wheat was longer at temperatures below 33/288C (Tashiro and
Wardlaw, 1989). The seed-ﬁll duration of a japonica rice cultivar did not change
between 33/28 and 21/168C while that of sorghum more than doubled over the
same range (Chowdhury and Wardlaw, 1978). The effect of shorter seed-ﬁll
Figure 1 The relationship between temperature and seed-ﬁll duration for several crop species.
Seed-ﬁll duration was estimated by the effective ﬁlling period and data from each source was averaged
across genotypes, years or experiments where appropriate. The regression was signiﬁcant at
P , 0:001: Maize—Tollenaar and Bruulsema (1988) and Wilhem et al. (1999); wheat—Vos (1981),
Tashiro and Wardlaw (1989), Hunt et al. (1991) and Wardlaw and Moncur (1995); rice—Fujita et al.
(1984) and Tashiro and Wardlaw (1989); sunﬂower—Chimenti et al. (2001).
durations at higher temperatures may be mitigated by an increase in seed growth
rate (Chowdhury and Wardlaw, 1978; Spiertz, 1978; Wardlaw and Moncur,
1995) that reduces or prevents changes in ﬁnal seed size. Comparisons of relative
temperature effects among experiments may be misleading because of variation
in techniques, but comparisons in the same experiment suggest that species
differences exist and are important.
The response to temperature during seed ﬁlling may not be the same as during
other reproductive growth stages. The simulation model CROPGRO uses a ﬂat
temperature response between 25 and 358C for all reproductive growth stages
in soybean, but the response to temperatures below 258C is less during seed
ﬁlling than during early reproductive growth (Boote et al., 1998). The rate of
development at 108C, relative to the maximum rate at 308C, was 0.2 during
early reproductive development compared with 0.8 during seed ﬁlling. Including
this differential response greatly improved the model’s predictive ability in
B. WATER STRESS
Water stress during seed development shortens the ﬁlling period and reduces
yield of many crop species, including wheat (Gallagher et al., 1976; Brooks et al.,
1982; Nicolas et al., 1984; Yang et al., 2000; Ahmadi and Baker, 2001), rice