Tải bản đầy đủ - 0trang
III. Biomass Production and Partitioning
MEDITERRANEAN WHEAT YIELD INCREASES
Hucl and Baker, 1987). Up to the 1970s, biomass production had not increased
for wheats bred by The International Maize and Wheat Improvement Centre
(CIMMYT), but more recent yield increases in cultivars bred under irrigation
have been associated with increased biomass under water-limited situations
(Evans, 1987). Also, Hucl and Baker (1987) found a positive correlation between GY and biomass production when comparing old and modem Canadian
Figure 4 illustrates changes in GY, biomass, HI, and maturity with the year
of release of Western Australian wheats. This figure is based on the data of Perry
and D’Antuono (1989), but also includes three new cultivars not included in their
study. The inclusion of these latest cultivars indicates that the rate of increase in
GY between 1965 and 1990 is greater than in the preceding 100 years, and the
rate of biomass increase associated with breeding is small (Fig. 4b). Most of the
GY increase can be attributed to increased HI (Fig. 4c), and the duration from
sowing to anthesis has decreased with selection for yield (Fig. 4d). This and
other studies (CIMMYT, 1991; Slafer et al., 1993) also demonstrate that modern
cultivars outperform older cultivars even in dry environments with low yield
Several authors (Donald and Hamblin, 1976; Richards, 1987; Turner and
Nicholas, 1987) suggest future wheat yields can be increased by increasing biomass production. Certainly, there is the potential for increased biomass production in mediterranean environments in some circumstances; barley is capable of
producing more biomass and grain than wheat using the same amount of water
(Siddique et al., 1989b; Lopez-Castaneda, 1992; Gregory et a!., 1992; Simpson
and Siddique, 1993). However, as with increased biomass caused by agronomic
manipulation, genetically increased biomass is not always translated into increased GY.
Where lack of water is the major limitation to growth, it appears that it will
be difficult to increase the biomass production of wheat significantly, especially
when soil water is completely exhausted at maturity. Under these circumstances,
higher biomass will be translated into higher yield when rainfall is used more
efficiently for photosynthesis and it may be difficult to improve these fundamental physiological processes with conventional breeding methods. We will discuss
in more detail the potential for increasing water-use and radiation-use efficiencies
later and deal with assimilate partitioning first.
The genetics of HI is probably more easily modified than biomass production,
but given the nature of moisture stress during grain filling, large increases in HI
are unlikely in mediterranean environments (Siddique et al., 1989b; Hadjishristodoulou, 1991). Aspects of the genetic, physiological, and environmental regulation of partitioning of assimilates were reviewed by Snyder and Carlson (1984),
Gifford er al. (1984), and Wardlaw (1990). Germination, ear initiation, terminal
S . P. LOSS AND K. H. M. SIDDIQUE
o o o p
Year of release
Figure 4 Mean cultivar grain yield, biomass, harvest index and maturity score of wheat cultivars released between 1860 and 1990 in Western Australia when grown under the same conditions.
Maturity score is the time from sowing to anthesis relative to Gamenya = 100, which takes about
110 days. Data from Perry and D'Antuono (1989) for cultivars up until 1979 (28 experiments each)
and from Siddique et al. (1989a.b). Regan et al. (1992), and Loss et al. (1989) for the three latest
cultivars (five experiments each).
spikelet, and anthesis act as physiological switches for the allocation of assimilates to different organs of the plant, hence phenology (i.e., the duration of each
of these development phases) and assimilate partitioning are closely related.
MEDITERRANEAN WHEAT YIELD INCREASES
Change in phenology is the single most important factor that accounts for
increased wheat yields in Australia (Perry and D' Antuono, 1989; Richards,
1991). Early Australian pioneers and wheat breeders recognized that matching
the crop life cycle to the length of the growing season is one of the most important factors influencing crop growth and yield. As we will discuss in more detail,
changes in phenology have had secondary effects on assimilate partitioning, pattern of water use, and other traits. Several reviews (Kirby and Appleyard, 1987;
Simmons, 1987; Hay and Kirby, 1991) have detailed the current understanding
of wheat development, and here we outline only the main features applicable to
The switches from one stage of development to the next are determined primarily by genes sensitive to photoperiod, and both high and low temperatures.
In general, there has been a trend to select for less sensitivity to photoperiod and
vernalization, especially in mediterranean environments, thereby advancing development and reducing the time to reach anthesis (Hake and Weir, 1970; Austin
et a l . , 1980; Davidson et al., 1985; Perry et al., 1987; Cox et a l . , 1988; Van
Oosterom and Acevedo, 1992). Of the regions that experience mediterranean
climates, wheat phenology has been described in the most detail in Western
Australia. In several comprehensive studies, Kirby and Perry (1987), Kirby et
al. (1989), Siddique e t a l . (1989a,b), and Loss e t a l . (1989) observed the development pattern of old and modern wheats bred in Western Australia, and illustrated how the life cycle of wheat has changed with selection for yield. Hence,
we give frequent examples from these studies.
1. Vegetative Development
The rates of leaf initiation and emergence in wheat are relatively constant
when plotted against thermal time (i.e. accumulated temperature, as defined by
Weir et a l . , 1984), although photoperiod and cultivar can have a small effect on
these rates. The rate of primordia initiation is about one every 50"Cd (degree
days, above a 0°C base), and the rate of leaf emergence is about one every
100"Cd (Kirby and Perry 1987). Tillers are initiated in the axils of the leaves
and, if conditions are suitable, the first tiller appears after 2.5-3 leaves have
emerged. Subsequent tillers appear at intervals equal to about one phyllochron
(the interval in thermal time between the appearance of one leaf and the next).
Ear initiation signals the end of vegetative development and the start of reproductive development.
In general, modern Australian cultivars have faster rates of vegetative development than old cultivars, including faster rates of leaf appearance, shorter durations of vegetative growth, fewer leaves, and, hence, fewer tillers (Kirby et
S. P. LOSS AND K. H. M. SIDDIQUE
Thermal time from sowing (OCd)
Figure 5 Duration between sowing, double ridge (DR), terminal spikelet (TS), anthesis (A),
and physiological maturity (PM) for Western Australian wheats (Kulin. Gamenya, and Purple Straw)
grown at Perth under irrigation. Year of release is indicated. Data from Kirby ef al. (1989) and Loss
er al. (1989).
al., 1989; Siddique et al., 1989a,b). For example, the phyllochron intervals
varied between 97 and 126"Cd for Western Australian wheats, and a modern
barley cultivar had a phyllochron interval of 84"Cd. The longest duration of
vegetative development was shown by the cultivar Purple Straw, released in the
1860s, which reached double ridge 958"Cd after sowing (Fig. 5). In contrast,
the cultivar Kulin, released in 1986, reached double ridge 424"Cd after sowing,
and, consequently, Kulin produced only 8 leaves on the main stem while Purple
Straw produced 14.
Roots have been less well researched than shoots because of the difficulties in
root collection and measurement, hence our understanding of root growth and
development is less complete than that for the shoot, particularly in mediterranean environments. This is not to say that roots are less important. In fact, root
growth is an important component of the adaptation of wheat to dryland environments. The relationships between root and shoot development have been described by Klepper et al. (1984) and were recently reviewed by Klepper (1992).
Root growth, with reference to mediterranean environments, is described later
in this article.
2. Ear Initiation
Ear initiation begins with spikelet initiation, progresses during a period of leaf
growth, and ends at terminal spikelet formation. Double ridge is easier to recognize than ear initiation, and for practical purposes these have been considered
The rate of spikelet initiation is faster than that of leaf initiation (Kirby and
Perry, 1987). Modern wheats have faster rates of spikelet initiation than do old
cultivars; one every 12"Cd for Kulin and 33"Cd for Purple Straw (Kirby et al.,
1989; Siddique et al., 1989b). Old cultivars have some vernalization requirement
MEDITERRANEAN WHEAT YIELD INCREASES
and when grown in areas with effective vernalizing temperatures, they have
shorter durations between ridge and terminal spikelet than do modern cultivars-238"Cd for Kulin and 167"Cd for Purple Straw (Kirby et al., 1989). However, in warm areas with similar photoperiods, the vernalization requirements of
the old cultivars are met more slowly and their duration between double ridge
and terminal spikelet is extended, whereas the duration in modern cultivars is
decreased (Fig. 5).
Bingham (1969) suggested extending the period of ear development to increase sink capacity of the grain, and, interestingly, this has occurred with selection for yield in Western Australia. The period between double ridge and anthesis
was 184"Cd longer in Kulin than in Purple Straw when grown in cool parts of
Western Australia (Siddique et al., 1989b). However, warm temperatures during
winter can cause rapid rates of development in wheats with little or no vernalization requirements and this results in reduced numbers of spikelets and number of
grains per spike (Warrington et al., 1977; Shpiler and Blum, 1986).
3. Floret Initiation and Stem Elongation
Floret development starts just before terminal spikelet formation, and coincides with the beginning of stem internode elongation. During this critical stage
of development, the potential number of grains and the yield potential of the crop
are determined, while there is an overlap of leaf, stem, root, and ear growth.
Under favourable conditions, the central spikelet of a developing ear can produce
up to 10 floret primordia, but only two to four survive and set grain in mediterranean environments (Siddique et al., 1989a; Slafer and Andrade, 1993).
The number of tillers reaches a maximum at terminal spikelet and declines
until anthesis. While old cultivars produce many more tillers than do modern
cultivars, tiller survival is much lower-35% in old and 5 1 % in modern cultivars
(Siddique et al., 1989b). The duration between terminal spikelet and anthesis of
modern cultivars was shorter than in old cultivars (Kirby et al., 1989)-825"Cd
for Kulin and 927"Cd for Purple Straw (Fig. 5 ) .
Anthesis signals the end of vegetative growth and the start of grain filling, and
its timing can have a large effect on cereal yields in mediterranean environments.
The timing of anthesis is particularly important for determinate plants, such as
cereals, because they only have a single opportunity for producing grain, as
opposed to indeterminate plants, which are able to produce flowers over a considerable period of time. The time of anthesis that produces the highest longterm yield is a compromise between sowing time and the risks of frost, low
biomass production, disease, high temperatures, and drought during grain filling.
Several studies have demonstrated that cereal yields increase when anthesis is
S. P. LOSS AND K. H. M. SIDDIQUE
advanced because of decreased high temperature and moisture stresses during
grain filling (Fischer and Kohn, 1966; Woodruff and Tonks, 1983; Kerr et ul.,
1991). In addition, metabolic energy is required for the storage and retranslocation of assimilates (Geiger and Fondy, 1980), and once the minimum required
structures of the plant have been produced, that is, the stem and leaves; then it
is more efficient to partition growth directly into ears and grain, rather than
producing additional vegetative growth and retranslocating the assimilates to the
grain at some later stage. Anthesis has been advanced by genetic means and by
the adoption of sowing very soon after the first autumn rains, when temperatures
are warmer than in winter, hence improving early growth.
Unlike ear initiation or terminal spikelet development, anthesis is easily visible and there has been a conscious selection for early anthesis in many environments (Halse and Weir, 1970; Austin et al., 1980; Davidson et ul., 1985; Perry
and D'Antuono, 1989; Cox et al., 1988; Van Oosterom and Acevedo, 1992).
Consequently, the duration between sowing and anthesis has decreased considerably (Fig. 5). In the study of Loss er al. (1989), Purple Straw flowered after
2132"Cd, whereas Kulin flowered after 1416"Cd.
Early anthesis can also have detrimental effects. Unfortunately, during the
period after emergence, the ear is very susceptible to frost damage, and as discussed earlier, frosts can cause devastating yield losses. Rapid development may
also reduce the amount of biomass produced at anthesis, the number of sites for
grain filling, and, hence, potential yield (Fischer, 1979). As was shown for sunflowers in Spain (Fereres et al., 1986) and sorghum in Texas (Blum and Arkin,
1984), very early-maturing plants may have restricted rooting depth and water
use. In some early-sown crops that flower early, the high temperatures during
vegetative growth increase the crop's susceptibility to diseases, especially on the
early-emerging flag leaf (Wilson, 1989).
Breeding for rapid development, that is, less photoperiod sensitivity and less
vernalization requirement, has caused more variation in date of anthesis. In individual years and at locations where the optimum period of anthesis is very
narrow, small variations in temperature can change crop development such that
the crop reaches anthesis beyond the optimum time. For example, in Western
Australia, farmers should sow early when the first autumn rains commence early
to make use of the available moisture and warm temperatures (Perry et af., 1989;
Kerr et al., 1991); however, there is a need for midseason cultivars that reach
anthesis during the optimum period from these early sowing times, particularly
in years when winter temperatures are above average. This could be achieved by
incorporating a small vernalization requirement into existing cultivars (Anderson
and Smith, 1990). Cultivars with a small vernalization requirement are less affected by temperature variations than are cultivars that require no vernalization,
because vernalization prevents rapid pre-ear initiation development should an
early break of the season be followed by a warm winter (Loss et al., 1990; Van
MEDITERRANEAN WHEAT YIELD INCREASES
Oosterom and Acevedo, 1992). Ludlow and Muchow (1990) also suggested that
a greater sensitivity to photoperiod will overcome the effect on the timing of
anthesis of year-to-year variations in temperature. With a large photoperiod sensitivity, anthesis is mainly triggered by daylength at a particular time of the year,
irrespective of temperature.
In contrast, yields may be increased by very early anthesis in dry areas with a
low risk of frost, especially in seasons when autumn rains are delayed and sowing is later than average. Very early anthesis, that is, less than 95 days from
sowing (<1 100-130O0Cd), may be appropriate for areas of North Africa (Van
Oosterom and Acevedo, 1992) and Western Australia.
5. Grain Development
The development of wheat culminates with the formation of grain. Grain filling can be divided into three phases. After anthesis there is a short period of
exponential growth, sometimes referred to as the lag phase, during which time
the cells of the endosperm divide rapidly and the potential size of the grain is
determined. During the second phase, starch is deposited in the endosperm and
the rate of growth is constant when expressed as thermal time. The final phase
begins when lipids are deposited in the phloem strands supplying the grain and
the growth rate declines until maximum grain weight is achieved.
The process of grain growth can be considered as two components-rate,
which is reflected in the rate of biochemical reactions involved in the synthesis
of starch and protein, and duration, which is a reflection of the developmental
program of the grain (Jenner et al., 1991). Loss et al. (1989) and Austin et al.
(1989) studied grain growth and development of old and modern wheats under
irrigation in a mediterranean and a temperate climate, respectively. Cultivars
with a short duration from sowing to anthesis, which was the case for most
modem cultivars, also had a long duration of grain growth, i.e., about 800°Cd,
or 200"Cd longer than the oldest cultivars (Fig. 5). This may explain why modern cultivars are more able to exploit seasons and environments where conditions
in spring are mild and grain filling is not terminated by drought and high temperatures. Modern cultivars that reached anthesis quickly also had shorter lag
phases than did old cultivars-6% of the duration of grain growth for Kulin and
18% for Purple Straw. We will discuss changes in the rate of grain growth later.
Breeders have changed the structure of cereals considerably, both indirectly
through changes in phenology and directly through the introduction of dwarfing
S. P. LOSS AND K. H. M. SIDDIQUE
1. Leaves and Tillers
One path to improved wheat yields is to increase HI by decreasing the proportion of biomass in leaves and tillers (Donald and Hamblin, 1976). Selection of
wheats for Western Australia and southwestern Iran has favored cultivars that
produce fewer main stem leaves and fewer tillers than the old cultivars (Siddique
ef al., 1989b; Ehdaie and Waines, 1989). Modern cultivars produce only primary tillers associated with the first two or three leaves. These tillers have three
or more of their own leaves, their own nodal roots, and they are largely independent of the subtending leaf for assimilate supply, hence their survival rate is high,
about 50%. Older wheats produce many more tillers and subtillers than do modern cultivars, of which 35% survive to produce grain (Siddique et al., 1989b).
Given the inefficiencies in retranslocation (Geiger and Fondy, 1980), tiller death
represents a net loss of assimilate by the plant and increases the water used before
anthesis. Siddique el al. (1989b) suggested that additional increases in yield in
mediterranean environments may result from further decreases in tiller number
and increases in tiller survival, particularly in low-rainfall areas where biomass
production is low. They acknowledged that this may also reduce the ability of
the plant to take advantage of years when conditions are favorable and tiller
survival is inherently high.
The single mainstem wheat, or uniculm, was first proposed by Donald ( 1 968)
as an important characteristic for dry environments. Islam and Sedgley (1981)
found that wheats that were surgically restricted to two shoots per plant used
more water after anthesis and produced greater GY s than did unrestrictedtillering wheats. Contrasting results were observed by Marshall and Boyd
(1985), who found that two Israeli biculms, despite having larger ears, yielded
25% less biomass and 30% less grain than conventional cultivars from Western
Australia. In addition to the differences in rainfall and soil types between the two
studies, the comparison within the Marshall and Boyd (1985) study was probably
complicated by the poor adaptation of the Israeli cultivars outside of Israel.
More recently, Whan et al. (1989) and Yunusa and Sedgley (1992) reported
no advantage of limited-tillering wheat breeding lines in water use or GY under
dryland conditions in Western Australia. Reduced tillering was also of no advantage in the semiarid wheat-growing regions of western Canada (Hucl and Baker,
1991) or for barley in South Australia (McDonald, 1990). In these cases, the
lack of yield increases with reduced-tillering wheats is probably related to their
larger leaves and to the absence of changes in leaf area index or biomass when
compared to conventional wheats (Richards, 1988). In addition, reducedtillering wheats tend to have other inefficient assimilate partitioning, that is, large
specific leaf weights, high stem and ear densities, and a high proportion of the
ear as chaff.
Turner and Nicolas (1987) proposed that rapid, vigorous seedling growth
should be advantageous on coarse-textured soils in water-limiting environments
WHEAT YIELD INCREASES
because of more efficient use of water. Seedling growth has important consequences for the pattern of water use and is discussed later in this article.
Root growth is strongly influenced by moisture and nutrient availability,
soil type, and cultivation (Hamblin et al., 1990). During seedling and tillering
growth, more assimilate is usually partitioned to roots than to shoots, but after
anthesis root growth is reduced, and when grown under favorable conditions,
roots frequently account for as little as 10%of the total crop biomass at maturity
(Lupton et al., 1974). Under drought conditions, however, roots may comprise
as much as 60% of the total crop biomass at maturity (Gregory et al., 1984;
Hamblin et al., 1990; Siddique et al., 1990b).
Hamblin and Tennant (1987) argue that rooting depth or rate of root elongation
is a better selection criteria for maximizing water uptake than either root length,
weight, or density in the mediterranean environment of Western Australia. Cereal root densities in the top 30 cm of soil are very high, compared with grain
legumes (Gregory, 1988), and this has been interpreted as evidence that less
roots in the surface soil may increase rooting depth and cereal yields (Richards,
1991); however, differences in root morphology and physiology are also important. Hamblin and Tennant (1987) also measured greater root lengths for cereals
when compared to grain legumes, but water uptake was better correlated with
rooting depth than with total root length. Water uptake per unit root length was
greater in the grain legumes than in the cereals, probably due to lower axial
resistances in the xylem vessels of the grain legumes. Smucker (1984) suggests
that cereals adopt a “conservative rooting strategy,” typified by an extensive root
morphology that uses water slowly. In contrast, legumes tend to have an “opportunistic rooting strategy,” which is characterized by a less extensive root
system that uses water rapidly.
Cereals have evolved from a predominantly nodal to a predominantly seminal
root system, particularly when grown at high densities (MacKey, 1986). Seminals develop earlier and deeper, and are finer and more efficient at water uptake
per unit dry weight than are nodal roots (Passioura, 1976). Seminal roots also
have a higher resistance to water flow and tend to conserve soil water more than
nodal roots. Producing more roots, particularly deeper roots to obtain more
water, appears to be a logical drought avoidance mechanism; however, roots are
a major sink for assimilates, requiring twice as much assimilate to produce the
same amount of biomass as shoots (Passioura, 1983). In cases where a small
amount of water is stored in the subsoil, the cost of producing deep roots to
obtain this water may be less than the extra assimilate that can be produced from
the additional transpiration and photosynthesis.
In m e d i t e ~ n environments,
soil moisture is almost always exhausted at
maturity (Cooper et al., 1987; Siddique et al., 1990b; Gregory et al., 1992),
S. P. LOSS AND K. H. M. SIDDIQUE
and cultivars that produce less roots, particularly in the top soil, may be at an
advantage. In fact, this is what has occurred with selection for yield in Western
Australia (Siddique et al., 1990a,b). Modern wheats produce less root dry matter
and lower root: shoot ratios than do old wheats, which probably relates to their
earlier ear sink development (double ridge), fewer tillers, and fewer adventitious
roots associated with the tillers. Patterns of root and shoot growth for old and
modern wheats are illustrated in Fig. 6.
Modem cultivars have root densities of 10 cm ~ m in-the~ top 10 cm of soil,
about half the density of the old wheats. Passioura (1983) estimated that root
densities of 0.5 cm ~ m are- adequate
for removing all the water stored in soils,
although higher densities may be required for nutrient uptake. In the past, roots
for nutrient uptake were less important because nutrient uptake could be improved by increasing fertilizer application, but, recently, more economical use
of fertilizer and reduced groundwater pollution from fertilizer leaching are considered desirable.
MacKey (1973) claimed that because root growth tends to mirror shoot growth,
semidwarf cereals may have reduced root growth when compared to tall cereals;
however, the studies of Siddique et al. (1990b) and Holbrook and Welsh (1980)
demonstrate this is not the case in dry environments. The growth of shoots by
modern semidwarf and old tall cultivars is similar, and modern cultivars produce
deep roots earlier than do old cultivars (Siddique et al., 1990b). By anthesis, the
roots of old and modern wheats reach the same depth. MacKey (1986) suggested
that root growth in modern wheats is reduced soon after anthesis because they
Root dry matter (g.m-2)
Figure 6 Patterns of root and shoot growth for Western Australian wheats grown at Merredin.
Terminal spikelet (TS)and anthesis (A) are indicated with arrows. Reproduced with permission from
Siddique eta!. (19Wa).
MEDITERRANEAN WHEAT YIELD INCREASES
have fewer lower leaves that provide the roots with assimilates, compared to old
Although advocated by Bums (1980), Passioura (1983), and Richards (1991),
it is not known if yields can be increased by further reductions in the growth of
roots in the surface soil. Using a simulation model, Miglietta et al. (1987) predicted that for a silty clay-loam, increasing rooting depth below 80 cm increased
water uptake early in the season but left little stored in the soil for use later in
the season, hence yields were reduced.
Techniques for measuring rooting depth are unsuitable for screening large
numbers of breeding lines, and we are aware of only one example where rooting
depth has been successfully incorporated into a wheat breeding program. In the
semiarid region of western Canada, Hurd et al. ( I 972) selected parents for deep
and prolific root systems in sloping plastic boxes and produced a deep-rooted
and high-yielding line that was later released as a commercial cultivar. Using
similar techniques, others have observed a wide variation in wheat root growth
(Derera el al., 1969; MacKey, 1983; Sharma and Lafever, 1992). Other techniques for examining roots, such as hydroponics, herbicide placement at various
soil depths, and minirhizotrons, have also been examined (Gregory, 1989), but
these are also unsuitable for selection in routine breeding programs.
3. Grain Growth
Temperatures above 30°C are common during grain filling in mediterranean
environments, causing an increase in the rate of grain growth. However, this
increased rate does not compensate for the reduction in duration of grain filling
(Sofield et al., 1977; Wardlaw et al., 1980, 1989) and, consequently, grain
sizes are smaller in mediterranean than in temperate environments. In fact, selection for yield in Australia has favored cultivars with many more grains but
smaller grain weights (Perry and D’Antuono, 1989). Cultivars that are able to
fill their grain quickly may reach physiological maturity before moisture stress
limits grain growth (Bruckner and Frohberg, 1987), and in areas of high frost
risk, cultivars with delayed anthesis and fast grain growth rates reduce the risk
of frost damage without increasing the risk of moisture stress during grain filling
(Loss et al., 1989).
The proportion of GY that is derived from assimilates produced after anthesis
varies from 70 to 95% depending on the degree of moisture stress (Rawson and
Evans, 1971; Austin et al., 1977; Bidinger et al., 1977; Pheloung and Siddique,
1991; Kobata et al., 1992). Although modem cultivars have fewer leaves, the
rate of grain growth is not reduced because most of the assimilate used for grain
growth is produced by the upper canopy. These assimilates are derived mainly
from the spike, the flag leaf, and its sheath (Austin and Jones, 1975; Rawson et
al., 1983), and it has long been recognized that awns can also substantially