Tải bản đầy đủ - 0 (trang)
II. Heterosis in Rice and Wheat

II. Heterosis in Rice and Wheat

Tải bản đầy đủ - 0trang

148



S. S. VIRMANI AND IAN B. EDWARDS



small number of crosses were evaluaw, parental selection was not necessarily

designed to maximhe heterosis;small populations were space planted either in

the field or in greenhouses; noncommercial and unproductive varieties were

frequently used (thereby eliminating the opportunity to evaluate standard heterosis); and the effects on height, maturity, and yield components were measured

more often than heterosis for grain yield. Despite these limitations, the levels of

heterosis have been high in certain cross combinations. However, Murayama et

al. (1974)provided evidence that heterosis in rice is not influenced by plant

spacing and soil fertility. In wheat, Jost and Glatki-Jost (1976)noted that excep

tional hybrids can produce more tillers per unit area than inbreds, regardless of

the seeding rate.

Suggestions of exploiting heterosis commercially by developing F, rice

hybrids have been made from time to time (Stansel and Craigmiles, 1966;Shinjyo and Omura, 1966a,b; Yuan, 1966, 1972;Craigmiles er al., 1968;Huang,

1970; Watanabe, 1971; Athwal and Virmani, 1972; Carnahan et al., 1972;

Swaminathan et al., 1972;Baldi, 1976). However, difficulties in hybrid seed

production discouraged most of the researchers from continuing their efforts, the

notable exceptions being Chinese scientists (Yuan, 1966, 1972).

In wheat, siflicant hybrid advantages have been measured in some instances, while other studies have reported no hybrid advantage (Kronstad and

Foote, 1964;Larrea,1966;Brown etal.. 1966;Briggle ef al., 1967;Fonseca and

Patterson, 1968; Livers and Heyne, 1968; Wells and Lay, 1970; Singh and

Singh, 1971;Bitzer and Fu, 1972;Allan, 1973;Widner and Lebsock, 1973;Jost

and Glatki-Jost, 1976; Yadav and Murty, 1976;Jost ef al., 1976b;Hughes and

Bodden, 1978;Cregan and Busch, 1978;Mihaljev, 1980;Jost and Jost, 1980;

Bailey et al., 1980;Wilson et al., 1980). Livers and Heyne (1968)reported on a

comprehensive 4-year study to determine hybrid vigor by intercrossing 9 varieties of well-adapted winter wheats at Hayes, Kansas. The 36 hybrids collectively exceeded all varieties by 20,37,37,and 35% in the 4 years (1964-1967),

respectively, with an average hybrid superiority of 32%. The best hybrid was

consistently better than the best variety for the area. Similar results were obtained

when 10 hybrids were compared with leading varieties at three locations (Livers

and Heyne, 1966).

When wheat hybrids were made using cytoplasmic male sterility-fertility

restoration systems and field tested at optimum population rates, the results were

less favorable than those obtained from hand-produced hybrids (Allan, 1973;

Hayward, 1975; Johnson, 1977, 1978; Edwards er al., 1980). Allan (1973)

found high-parent hetemis ranging from 23 to 113% among soft white winter

wheat hybrids grown in the state of Washington, but the results were highly sitespecific; when averaged over 4 locations, they indicated that no hybrid had

outyielded the high parent. Jost and Milohnic (1975)tested 5 hybrids in Yugoslavia and, in this small sample, only 1 hybrid showed high-parent heterosis.



HYBRID RICE AND WHEAT



149



Similarly, Hayward (1975) reported on a 1970-1971 study of 39 hard red winter

wheat hybrids and 6 high-yielding check cultivars in Kansas; he found only 1

hybrid that equalled the yield of the highest check. However, Hayward did show

significant improvements among later hybrids tested in 1973-1974; the mean

yield of 4 hybrids over three locations was 19%greater than the mean of 8 check

cultivars, and the best hybrid yielded 13.7%more than the top cultivar. Some

possible reasons for the relatively poor performance of CMS-produced hybrids in

the early studies are inadequate fertility restoration, adverse effects of T. rimopheevi cytoplasm, heavy selection pressure for restoration ability with less

emphasis on agronomic performance during restorer development, and limited

testing of different hybrid combinations.

Most of the hard red winter wheat hybrids evaluated have shown more specific

adaptability to certain regions and winter-hardiness zones than common check

cultivars. Johnson (1977, 1978) reported on tests conducted at 11 sites in five

states (Texas, Oklahoma, Kansas, Colorado, and Nebraska) during 1975-1979

and 1976-1977. Of the 15 hybrids tested in 1976, the leading hybrid produced

70 kg/ha less than the check cultivar ‘Centurk.’ In 1977,16 hybrids averaged 70

kg/ha less than Centurk, although the best hybrid exceeded Centurk by 130

kg/ha. Johnson concluded that the hybrids were not sufficiently superior to

justify their use over the best available varieties. Published reports of spring

wheat hybrid evaluation have been more limited compared with winter wheats.

Edwards et al. (1980) reported on 1978 and 1979 tests with a series of spring

wheat hybrids. Although the levels of high-parent heterosis ranged up to 35%,

the top-yielding hybrids exhibited standard heterosis of only 10-14% above the

leading check cultivar ‘Era.’In summary, hybrid wheats based on the cytoplasmic-genetic system have not expressed as much heterosis as the early handproduced hybrids, and the yield advantages to date have not been sufficiently

great to justify widespread commercial production.

Experiments on heterosis in rice conducted at Davis, California (Rutger and

Shinjyo, 1980) indicated significant yield superiority of 11 of 153 rice hybrids

over the best check variety. Standard heterosis ranged from 16 to 63% and

averaged 41%. Hybrid corn seed producers in the U.S. maintained that this

frequency (11 of 153 combinations) and degree of standard heterosis (41%)

would make the prospects of hybrid rice exciting if sufficient hybrid seed could

be produced (Rutger and Shinjyo, 1980).

Studies conducted at the International Rice Research Institute in the Philippines during 1980-1981 have shown levels of as much as 73, 59, and 34% for

mid-parent, high-parent, and standard heterosis, respectively (Virmani et al.,

1982). Hand-crossed F, hybrids produced from elite breeding lines yielded up to

6.2 ton/ha compared with 5.0 ton/ha from the best check variety (‘IR42’)in the

wet season, and 10.4 ton/ha compared with 7.9 ton/ha (IR54) under irrigation

during the dry season. Of a total of 202 F, hybrids evaluated for yield during



150



S.



S. VIRMANI AND IAN B. EDWARDS



1980-1982, 63% showed positive high-parent heterosis (4-64.3%) and 50%

showed positive standard heterosis (0.1-46.4%). Yields of the F, were found to

be positively correlated with the parental mean. (r = 0.45**) and high-parent

values (r = 0.33**). The parents had been selected for high per se yield performance, diverse genetic background, and resistance to diseases and insects incorporated though conventional breeding, Saini et al. (1974) also observed positive

standard heterosis when the F, hybrids were derived from selected parents with

improved plant type. However, the association between parental yield per se and

F, hybrid performance may vary with the genetic background of the inbreds, and

Khaleque et al. (1977) found no association between parental and hybrid yield

performance.

The most comprehensive commercial utilization of heterosis in rice has been

that reported from the People’s Republic of China (Li, 1977; Lin, 1977; Lin and

Yuan, 1980); more than 12 hybrids were officially released prior to 1980 (Shen,

1980). Yields under large-scale production have exceeded the best conventionally bred varieties by 20-30%. Results from replicated yield trials are given

in Table I, and the data indicate that although the hybrids had fewer effective

panicles per square meter, they had significantly more filled grains per panicle

and larger seeds. The highest individual yield obtained from the F, hybrids was

12.8 tonslha, compared with 10.4 tons/ha from a conventionallybred variety (L.

P. Yuan, personal communication).

The major yield components in rice and wheat are number of panicles (or

spikes) per square meter, spikelet number per panicle (or spike), spikelet fertility

percentage, and 1000-grain weight. Significant positive mid-parent, high-parent,

and/or standard heterosis have been observed for one or more of these components in a number of rice crosses (Pillai, 1961; Namboodiri, 1963; Rao, 1965;

Dhulappanavar and Mensikai, 1967; Karunakaran, 1968; Carnahan et al., 1972;

Chang et al., 1973; Mohanty and Mohapatra, 1973; Saini and Kumar, 1973;

Sivasubramanian and Madhava Menon, 1973; Murayama et al., 1974; Saki er

al., 1974; Parmar, 1974; Paramsivan, 1975; Davis and Rutger, 1976; Mallick ef

al., 1978; Rutger and Shinjyo, 1980; Virmani et al., 1981, 1982). Virmani el al.

(1981) observed negative heterosis for panicle number per square meter, but in

combinations showing positive mid- and high-parent heterosis for yield this was

overcompensated by positive heterosis in spikelets per panicle. Most crosses

showing significant standard heterosis for yield have been found to show heterosis for more than one component (Saini et al., 1974; Mauya and Singh, 1978;

Virmani et al., 1981, 1982). Results obtained in China and at the IRRI indicate

that heterotic F, combinations usually show an increased sink size through

increases in spikelets per panicle, spikelet fertility percentage, and 1000-grain

weight.

In wheat, Livers and Heyne (1968) pointed out that each yield component was

important but that no single one was predominant in determining yield. Their



HYBRID RICE AND WHEAT



151



Table I

Yield and Yield Components of Hybrid Rice Varieties in Regional Teats of 27 Sites in Hunan

Province, china0

Hybrid combinations

or check

1977

Wei You 6

Shan You 6

Nan You 6

Zhao You 6

Dong Ting Wan Xian (ck)

1979

Wei You 6

v20 x s

Tan You 4

Dong Ting Wan Xian (ck)



Yield

(tonsha)



Effective

panicledm2



Filled grains

per panicle



weight (g)



6.1

6.1

6.0

5.9

5.2



324

308

339

318

374



86.1

95.1

86.4

85.4

75.1



26.7

25.4

24.5

26.0

22.0



6.5

6.3

6.1

5.4



328

266

328

345



78.0

72.7

66.9

77.7



27.0

33.9

31.9

21.7



1OOO-grain



“Data from VimuCni ef d.(1981).



data showed that although top-yielding hybrids tend to have relatively high

values in all three components, good performance is possible with a low value

for any one component if the other two components have high values. Yieldcomponent compensation has been well documented in the literature (Donald,

1962; Bingham, 1967), and several workers have concluded that yield-component selection has limited value in breeding programs (Rasmusson and Camel,

1970; Fisher, 1975). However, kernel weight has been considered the most

independent yield component because it is the last component developed, and its

level of expression should not produce a compensating change in other components. In contrast, Sinha and Khanna (1975) hypothesized that heterosis in wheat

will have commercial utility only when yield per spike increases, because tiller

number per plant is strongly influenced by environment and can be manipulated

by seeding rate. Briggle et al. (1967) also noted that heterosis for tiller number

decreased as the population increased in b o a parents and hybrids. However, Jost

and Glatki-Jost (1976) found that exceptional hybrids had the capacity to produce

more tillers per unit area than inbreds irrespective of seeding rate, and Wilson et

al. (1980) found the 17% high-parent heterosis in full dwarf X semidwarf

hybrids to be primarily the result of an increase in spikes per unit area.

B. HETEROSISFOR OFHER PLANT CHARACTERS



In both rice and wheat, a number of studies have shown that heterosis for plant

height is highly cross specific (Pillai, 1961; Namboodiri, 1963; Briggle et al.,



152



S. S. VIRMANI AND IAN



B. EDWARDS



1964;Rao, 1965; Dhulappanavar and Mensikai, 1967; Livers and Heyne, 1968;

Karunakaran, 1968; Amaya et al., 1972; Bitzer and Fu, 1972; Sivasubramanian

and Menon, 1973; Ingold, 1974; Paramsivan, 1975; Khaleque et al.. 1977;

Sreekumari et al., 1977; Mallick et al.. 1978; Wilson et al., 1980), and significant positive as well as negative heterosis has been reported. In rice, shorter

height of F, hybrids may be attributed to the higher photoperiod sensitivity of

some hybrids in comparison to their parents. In wheat, various studies have

shown F, hybrids to exceed the tall-parent value (Ingold, 1974), exceed the midparent value (Amaya et al., 1972), approximate the mid-parent value (Wilson et

al., 1980), and, in the case of “Olsen-dwarf” derivatives, approach the dwarfparent value (I. B. Edwards, personal observation). Because height is one expression of vigor that may lead to unfavorable grain/straw ratios and belowoptimum yields as a result of lodging, a number of hybrid programs are manipulating dwarfing genes in parents to obtain desirable height expression in the

hybrids.

A number of workers have observed the growth duration of rice hybrids to be

shorter than the mid-parent value and, in some cases, shorter than that of the

early parent (Dhulappanavar and Mensikai, 1967; Karunakaran, 1968; Chang et

al., 1973; Bardhan Roy et al., 1975; Khaleque et al., 1977; Mallick et al.,

1978). The dominance of Efgenes for the short basic vegetative phase (BVP) has

been pointed out by Chang et al. (1969). However, in subtemperate-to-temperate

China, most of the heterotic rice hybrids are later maturing than their parents (Lin

and Yuan, 1980). This apparent discrepancy requires further investigation.

Heading dates of wheat hybrids made on both normal and alien cytoplasms have

tended to be earlier than the mid-parent value (Livers and Heyne, 1968; Amaya

et al.. 1972; Bitzer and Fu, 1972; Wilson et al., 1980) and, in some combinations, to exceed the early parent value (Bitzer and Fu, 1972; Jost et al., 1976a;

Jost and Hayward, 1980). I. B. Edwards (Table 11) found different sets of spring

wheat hybrids, evaluated over a 5-year period, to consistently show dominance

or overdominance for earliness. The latter is a desirable trait in the northern

spring wheat region of the United States, because heat stress is frequently encountered during the early grain-filling period.

A number of researchers have emphasized the need to maintain a complementary balance between “source” (photosynthate supply) and “sink” (potential

grains) in cereals. Sinha and Khanna (1975) proposed that both source and sink

capacity should increase in order to improve yield, and Virmani et al. (1981,

1982) observed this phenomenon in F, rice hybrids derived from semidwarf

parents. Jennings (1967) found significant heterosis for vegetative growth to be

negatively associated with yield in hybrids derived from tall parents. In contrast,

Virmani et al. (1981, 1982) observed significant standard heterosis for both

vegetative growth and grain yield in certain hybrid combinations. Increases in



HYBRID RICE AND WHEAT



153



Table II

Average Hybrid Performance of Spring Wheat during a 5-Year Period“



Measurement

Days to 50% heading

Low parent

Hybrid

High parent

Height (in.)

Low parent

Hybrid

High parent

Yield (kgha)

Low parent

Hybrid

High parent



1978

(9)b



1979

(12)



1980

(13)



1981

(9)



1982

(32)



55.5

54.7

59.2



56.4

56.I

59.3



57.5

56.6

61.1



64.9

65.2

68.8



49.7

48.0

51.5



30.6

33.3

34.3



30.9

34.0

35.I



28.9

31.6

32.6



28.7

32.2

32.9



29.2

32.5

31.8



38.4

52.1

47.8



29.6

43.2

36.9



53.4

55.6

58.7



34.9

30.2

42.6



51.7

62.7

66.3



“Source: I. B. Edwards, Pioneer Hi-Bred International, Inc.

bNumber of hybrids evaluated with their parents.



grain yield in certain hybrid combinations have been attributed to a more efficient distribution of dry matter in the plant, and the harvest index (ratio of grain

weightkotal plant weight) has been examined by several workers (Sinha and

Khanna, 1975). Benson (1978) found the high yields and heterosis in four spring

wheat hybrids to result from both increased plant weight and harvest index. He

attempted to use harvest index as a screening technique for yield in spring wheat

hybrids. However, although a high correlation was shown between harvest index

and yield in conventional-sized plots (2.44 X 1.22 m), the harvest index of

small, single-row plots showed only a weak correlation with yield in conventional plots.

In rice research, heterosis has been observed for such traits as cold tolerance

(Sawada and Takahashi, 1977), salt tolerance (Akbar and Yabuno, 1975), photoperiod sensitivity, and rooting habit (Lin and Yuan, 1980). Chinese F, hybrids

showed heterosis for root penetration rate, depth and width of the rhizosphere,

number of adventitious roots per plant, and number of root fibrils (Anonymous,

1977; Lin and Yuan, 1980). Preliminary observations made at IRRI have also

indicated that some hybrids are superior to their parents at comparable growth

stages with regard to total root dry weight and root number, length, diameter,

and pulling force (O’Toole and Soemartono, 1981).

Superiority of F, rice hybrids in such physiological traits as photosynthetic

area, chlorophyll content per unit area, photosynthetic efficiency, and mitochondrial activity has been reported in China (Hunan Agricultural College Depart-



154



S. S. VIRMANI AND IAN B. EDWARDS



ment of Chemistry, 1977; Lin and Yuan, 1980) and elsewhere (McDonald et al.,

1971, 1974). In wheat, Sage and Hobson (1973) observed increased mitochondrial activity above the high-parent value in several mixtures, and found these to

be significantly correlated with the percentage yield heterosis of fully restored

hybrids grown at lower seed densities. Improvements in both plant type and

physiological efficiency would appear to be a logical consequence of hybrid

research in both rice and wheat.

C . COMBINING

ABILITY



The diallel analysis has been the major mating design used to estimate heterosis and the relative amounts of general combining ability (GCA) and specific

combining ability (SCA) in rice and wheat. Most wheat studies have revealed

that GCA is usually of greater relative importance for grain yield than is SCA

(Kronstad and Foote, 1964; Brown et al., 1966; Gyawali et al., 1968; Walton,

1971; Bitzer and Fu, 1972; Widner and Lebsock, 1973). All workers reported

significant GCA effects for grain yield, but significant SCA effects occurred only

when the experiments were space planted (Kronstad and Foote, 1964; Gyawali er

al., 1968; Yadav and Murty, 1976). The absence of SCA effects in competitive

growth conditions suggests that nonadditive genetic variance may not be well

expressed in wheat under these circumstances (Cregan and Busch, 1978). Widner and Lebsock (1973) evaluated a 10-parent diallel of genetically diverse

durum wheat lines and found highly significant GCA effects for grain yield,

tillers per unit area, kernels per spike, kernel weight, seedling vigor, maturity,

height, and lodging. Specific combining activity effects among F, values were

significant for kernel weight, seedling weight, and seedling vigor, suggesting

that maximum grain production may be attainable under a system that can exploit

both additive and nonadditive genetic effects. The largest levels of heterosis and

the highest yielding hybrids involved genetically diverse parents, and other

workers have concluded that in hybrid wheat the genetic diversity of the parents

is as important as their mean performance (Nettevich, 1968; Yadav and Murty,

1976).

The results of combining ability studies in rice have tended to be more variable

than those in wheat. The predominant role of additive effects was established for

all yield components except panicle number, which was affected by a certain

level of nonallelic interaction (Chang et al., 1973; Li, 1975). Several workers

have found high GCA effects in the parents to be associated with maximum SCA

effects and heterosis for yield in the resulting hybrids (Ranganathan et al., 1973;

Parmar, 1974; Maurya and Singh, 1977; Singh, 1977; Khaleque et al., 1977;

Rahman et af.,,1981). In contrast, a number of inheritance studies have suggested dominant gene action for yield and/or yield components (Wu, 1968a,b;

Chang, 1971; Sivasubramanian and Madava Menon, 1973; Shaalan el al.. 1975;



HYBRID RICE AND WHEAT



155



Singh and Nanda, 1976; Singh et af., 1979, 1980; Rahman et af., 1981). Chang

(1980) found 1-2 pairs of dominant genes to affect the expression of heterosis

for panicle and grain characteristics. Finally, some workers have suggested that

their results showed little relationship between combining ability effects and the

manifestation of heterosis in the corresponding hybrids (Mohanty and

Mohapatra, 1973; Parmar, 1974; Singh and Nanda, 1976; Maurya and Singh,

1977; Rao et al., 1980; Haque et al., 1981; Rahman et al., 1981). Clearly there

is a need for hybrid rice programs to investigate this subject further.

The question of inbreeding depression has received comparatively little attention in rice and wheat although this is of major significance in assessing the

merits of hybrid versus conventional breeding. Cregan and Busch (1978) studied

the F, ,F,-F, bulks, and F5 lines from an eight-parent spring wheat diallel at two

locations. The F, yields showed significant GCA and SCA mean squares. The

latter was attributed to additive X additive epistasis and, although it was present

in the F, progeny, it was less apparent in later generations. The significant F,

heterosis and SCA for yield, coupled with significant inbreeding depression

(0.23% yield reduction per 1% decrease in heterozygosity), indicated the possible desirability of F, hybrids to maximize yields. However, no F, hybrid significantly outyielded the best F5 line tested, and it was unclear whether yields would

be maximized by pure line or F,-hybrid development. The inbreeding depression

in these crosses between genetically related parents, although significant, was

substantially smaller (one-half to one-third) than that reported in maize. This

was attributed to the significantly less dominant genetic variance. Yadav and

Murty (1976) were able to show varying levels of inbreeding depression in their

eight-parent spring wheat diallel study. A high level of inbreeding depression

was associated with high heterotic effects in diverse crosses. It is evident that

hybrid programs should establish separate heterotic pools for male and female

inbred development, that the genetic relationships between these pools should be

minimized, and that further studies of the relationship between heterosis and

inbreeding depression should be conducted.



Ill. ADVANTAGES OF HYBRIDS OVER

CONVENTIONALLY BRED VARIETIES

Hybrid advantages are not simply a function of heterosis. Three factors affect

the end result: (1) breeding-method efficiency (a rate-of-progress factor), (2) the

negative or positive effects of the cytoplasmic male sterility-fertility restoration

system used to produce the hybrid, and (3) the inherent heterosis.

Although the levels of heterosis in rice and wheat are comparable to those

obtained in maize and sorghum, comparatively few studies have reported eco-



156



S. S. VIRMANI AND IAN



B. EDWARDS



nomically significant yield advantages over the best conventional varieties. The

most promising results on hybrid rice have come from China (Lin and Yuan,

1980), where F, hybrids outyielded conventionally bred varieties by 20-30%

under varying levels of on-farm management. Preliminary results from IRRI

indicated that leading hybrids had high-parent and standard heterosis of 0-64 and

0-4796, respectively.

In wheat, few studies have reported economically significant yield advantages

of F, hybrids over the best conventional varieties. However, much of the research on hybrid wheat has been directed at perfecting the genetic system. Only

since the early 1970s have the leading hybrid programs devoted more research

effort toward the agronomic improvement of male inbreds (restorer lies). The

identification of varieties potentially useful as female inbreds, their rapid incorporation into a male-sterile conversion program, and the maintenance of pure

seed have provided a management challenge to hybrid breeders. It is against this

background that one must assess the advantages of hybrids over conventionally

bred varieties. A strong variety breeding program is fundamental for the production of female inbreds; to this extent hybrids and varieties are both complementary and competitive.

When separate, large, and genetically diverse “pools” of male and female

inbreds are available to a hybrid breeding program, it is reasonable to assume

that consistently higher levels of heterosis will be obtained. The hybrid program

may have considerable advantage over conventional programs that frequently

suffer from inbreeding situations in which several of the top parents used in

crosses have varying degrees of genetic relationship. At this point, the rate of

progress factor in hybrid production becomes significant [i.e., the generations of

selections (usually F,-F,) in conventional breeding are bypassed and the testing

phase is immediate]. The hybrid,breeding approach can expedite the incorporation of dominant genes for resistance to major diseases and insects. For example,

IR26 is a rice restorer line containing dominant genes for resistance to the brown

planthopper and bacterial leaf blight; it has conferred resistance to a number of

hybrids (Shen, 1980). If resistance is conditioned by recessive genes, these

would have to be incorporated into both parents.

In wheat, hybrids offer an advantage for trait complementation of certain

quality factors. For example, a mixing time in dough development that is either

too long or too short is considered an undesirable trait. Results indicate that

mixing time in the hybrid is intermediate between that of the two parents, and

this would produce a desirable result in such a cross. More vigorous vegetative

growth, taller height, stronger root systems, and higher photoperiod sensitivity

are some of the traits already observed in F, rice hybrids compared with their

parents. These traits may aid in suppressing weed competition and enable

hybrids to adjust to varying water and nutrient regimes. Yap and Chang (1976)

repoM that hybrids performed better under dryland than under wetland

conditions.



Tài liệu bạn tìm kiếm đã sẵn sàng tải về

II. Heterosis in Rice and Wheat

Tải bản đầy đủ ngay(0 tr)

×