Ying Wang,Lubiao Zhang,Afif Nafisah,Linghua Zhu,Jianlong Xu*,Zhikang Li*

Institute of Crop Science,The National Facility for Crop Gene Resources and Genetic Improvement,Chinese Academy of Agricultural Sciences,Beijing 100081,China

1.Introduction

Rice (Oryza sativa L.),as the most important staple food,feeds about 50% of the world population [1].However,world rice production has to increase by at least 70% by 2050 in order to meet the demand of the population.Historically,at least 50%of the increases in rice productivity have resulted from development and wide adoption of new cultivars,which included benefits of the Green Revolution in the 1960s and hybrid rice technology from the late 1970s.Nowadays,it is a priority to improve yield potential in almost all rice breeding programs worldwide.Meanwhile,rice production is facing more and more challenges,such as water scarcity resulting from urban and industrial demands and pollution [2,3],dramatically declining arable lands and land degradation [4],and more frequent and dramatic climate changes from global warming[5–8].Of these,drought is the most important factor limiting rice yield in the rainfed areas that account for ～35%of the world rice area.Rice yield reductions from drought in rainfed areas range from 20 to 100%.Similarly,salt stress is the second most important abiotic stress limiting rice productivity,particularly in coastal areas and some inland rice fields.It is estimated that 20–50% of the irrigated rice lands are somewhat salt-affected [9].Frequently,drought goes hand in hand with salinity in many areas of Asia where irrigation is used to reduce soil salt in rice paddy fields.For instance,reduced fresh water in irrigation often induces secondary salinization and aggravates the effects of salinity.Alternatively,secondary salinization worsens the effects of drought on rice.

To achieve high yield (HY) and yield stability through breeding,breeders have to develop high yielding rice varieties with significantly improved tolerances to drought and salinity.Challenges then arise from the fact that HY,drought tolerance(DT)and salt tolerance(ST)are all complex traits controlled by polygenes,possible negative associations of rice DT or ST with HY,and different genetic and physiological mechanisms of the same traits at different developmental stages [10–12].In addition,selection of the right parental lines as donors for target traits has been difficult in real breeding programs.For instance,many rice landraces have good levels of DT and ST,but are low yielding[11].Genetic drag is another major concern to breeders when they are making decisions in choosing landraces as trait donors,particularly when the conventional pedigree breeding method is used[13].

While commonly used to improve single highly heritable traits,backcross (BC) breeding and strong phenotypic selection have been proven to be effective for improving single complex traits,particularly abiotic stress tolerances in rice [14–16].However,when aiming at improving multiple complex traits using phenotypic selection in a real BC breeding program,breeders are facing several important and tricky issues regarding what selection strategy should be used.This is particularly true when breeders have to deal with trait selection in two contrasting environments–the normal summer crop season(s)in the target environments(TEs)and short-day winter nurseries of the tropical climate in Hainan,in order to speed up the breeding process.Thus,it remains unclear to most breeders as to what traits or trait combinations should be selected in each of the breeding environments.In particular,in what order and what environments,should different target traits be selected to achieve the best overall genetic gain within the shortest time,when multiple complex target traits have to be improved.

In this study,we tried to answer these questions by presenting results from an effort for simultaneously improving HY,DT and ST of rice using introgression breeding.While we were able to develop many promising high yielding rice lines with significantly improved DT and/or ST,our results also provided some insights into the optimal selection strategy for improving multiple complex traits based solely on phenotypic selection.

2.Materials and methods

2.1.Parental lines and the BC breeding procedure

Huang-Hua-Zhan(HHZ),a high yielding indica variety from South China was used as the recurrent parent(RP)to cross with three donors,IR64(indica from IRRI),AT354(indica from Sri Lanka)and C418(japonica from Northeast China).The F1s were backcrossed to HHZ to produced 3 BC1F1populations,from each of which 20–25 random BC1F1plants were further backcrossed to HHZ to produce 20–25 BC2F1lines.The selfed seeds from all BC2F1plants from a single cross were bulk harvested,forming a single BC2F2population.The BC2F2populations were subjected to three different selection schemes for improving the target traits (ST,HY and DT),as described in Fig.1.

2.2.Trait selection

Three different selection schemes were adopted in this study.In the first selection scheme,each of the bulk BC2F2populations was screened for DT under natural drought stress conditions(the soil type is sandy yellow clay) at the CAAS experiment station in Hainan during the 2009–2010 dry season (DS).Seeds of the bulk BC2F2populations were sown in the nursery on November 20,2009,and 400 25-day old seedlings of each BC population were transplanted into a 40-row plot with one row of HHZ inserted every 20 rows as checks.The spacing was 20 cm × 17 cm.Drought stress was started by draining the field at the peak tillering stage 30 days after transplanting.One irrigation was applied to prevent death when drought stress was severe at 65 days after the stress started.At the maturity,19,29 and 33 plants that had obviously higher yield than HHZ were visually selected from the HHZ/IR64,HHZ/AT354 and HHZ/C418 populations.The 81 DT selected HHZ BC2F3introgression lines(ILs)and HHZ were progeny tested under both irrigation and drought stress conditions at the CAAS experimental station in Beijing in the 2010 summer.Seeds of each IL were sown into a seeding tray and 45 30-day old seedlings of each IL were transplanted into a 3-row plot with a spacing of 20 cm × 17 cm and two replications for each IL and HHZ under each water treatment.In the irrigated control,a 5 cm layer of standing water was maintained in the field until 10 days before harvest.The crop management was the standard one with two applications of fertilizers at a total rate of 120 N kg ha-1.Under the drought treatment,all ILs were evaluated in the greenhouse at the CAAS experimental station with the same experiment design.A terminal drought was initiated by withholding irrigation 30 days after transplanting until maturity,drought conditions that were very severe killing HHZ and some ILs.However,43 ILs survived the stress and were able to produce seeds.These included 12 ILs from the HHZ/IR64 population,23 ILs from the HHZ/AT354 population,and 8 ILs from the HHZ/C418 population(Fig.1).Heading date(HD,in days)were visually recorded for all plots when ≥50%of the plants in a plot had headed.After complete heading,the plant height(PH,in cm)was measured from the soil surface to the tip of the tallest panicle (awns excluded).At maturity,five representative plants in each plot were harvested by cutting the plants at the soil surface.The harvested plants were dried naturally for a month and then measured for panicle number per plant (PN),spikelet number per panicle(SNP),filled-grain number per panicle(FNP),spikelet fertility(SF,in%),thousand-grain weight(GW,in g)and grain yield per plant(GY,in g).

In the second selection scheme,seeds of the three bulk BC2F2populations were sown in the seedling nursery at the CAAS experimental station in Beijing on May 10,2010.On June 4,480 25-day old BC2F2seedlings from each population were transplanted into a 40-row plot with 3 rows of HHZ(the recipient)inserted in every 10 rows as the checks.The field was managed using standard practices under normal irrigated conditions.At maturity,high yielding(HY)plants were visually identified and harvested and dried naturally for 10 days prior to measuring grain yield.Plants with at least 10%higher yield than HHZ were selected,resulting in 26,16 and 22 HY plants from the HHZ/IR64,HHZ/AT354 and HHZ/C418 populations.

In the third selection scheme,the three BC2F2populations were subjected to strong selection for seedling ST in the screen-house of CAAS in the 2009 summer.In this screen,seeds of the BC2F2populations and RP were sterilized with 1%sodium hypochlorite solution for 10 min and rinsed well with distilled water,then germinated at 35 °C for 48 h.Two germinated seeds were sown in a hole in a thin styrofoam board with 130-holes in 13 rows and a nylon net bottom.The styrofoam board was floated on water up to the two-leaf stage,and then the styrofoam board with seedlings was transferred to a plastic box filled with Yoshida cultural solution [17] containing 140 mmol·L-1NaCl in the screen-house of CAAS in Beijing.The solution was changed every 5 days and the daily pH was maintained at 5.5.Each styrofoam board had 240 plants from each population plus one row of HHZ and IR29 (the salt sensitive check) placed in the middle as checks and each population comprised two boxes.In the screen-house,a 29/22 °C day/night temperature and minimum relative humidity of ～70% was maintained with humidifiers.Approximately 18 days after the salt treatment when HHZ were killed,57,49 and 56 surviving seedlings were selected from the HHZ/IR64,HHZ/AT354 and HHZ/C418 populations,and transferred to the field for seed production.In the 2010 summer,the selected ILs were progeny tested for ST using the same method in the phytotron with two replications for each IL.Individual plants of each IL and HHZ were evaluated score of salt toxicity of leaves(SST)10 days after salinization and for days of seedling survival(DSS)each day from seeding to death according to the standard evaluation system (SES) [18].Based on the progeny testing,25,28 and 29 ILs from the HHZ/IR64,HHZ/AT354 and HHZ/C418 populations,respectively,were confirmed to have significantly improved seedling ST compared to HHZ.

Fig.1-Breeding procedure for improving three complex traits(high yield,drought tolerance and salinity tolerance)of three Huang-Hua-Zhan(HHZ)BC2 populations(HHZ/IR64,HHZ/AT354 and HHZ/C418).Numbers in parentheses are the selected introgression lines(ILs)in each round of selection(see Section 2).HY:high yield;DT:drought tolerance;ST:salinity tolerance.1:HHZ/IR64;2:HHZ/AT354; 3:HHZ/C418.

2.3.Final evaluation of the ILs

In the 2010–2011(Nov.2010 to June 2011),all 189 BC2F4and BC2F5ILs obtained from the three selection schemes were evaluated in replicated field experiments for their yield traits under drought stress and normal irrigated conditions in Hainan.Seeds of each IL were sown into a seeding tray on Nov.25,and 30-day old seedlings of each IL were transplanted into a 3-row plot with a spacing of 20 cm × 17 cm.The plots were arranged in a random complete block design with two replications for each IL in each water treatment.In the drought treatment,the drought stress was started by draining the field at peak tillering 30 days after transplanting.But the climate of this season in Hainan was not normal with a lower average temperature than normal,resulting in prolonged growth duration.A one-time flush irrigation was applied on Mar.20 when the drought stress appeared to be very severe.In the normally irrigated control,everything was the same as in the drought stress experiment except a 5 cm layer of water was maintained in the field until 10 days before harvest.Days to heading (HD,in days) were recorded for all plots when ≥50% of the plants in a plot were completely headed.After heading,the plant height(PH,in cm)was measured from the soil surface to the tip of the tallest panicle(awns excluded).At maturity,five representative plants in each plot were harvested by cutting the plants at the soil surface.The harvested plants were sun-dried for 7 days and the dried plants were measured for panicle number per plant(PN),spikelet number per panicle (SNP),filled-grain number per panicle(FNP),spikelet fertility(SF,in%),thousand-grain weight(GW,in g)and grain yield per plant(GY,in g).

2.4.Data analyses

ANOVA was performed to evaluate trait differences between the water treatments (T),among different ILs (G) within each water treatment,among different ILs within each population,between ILs from different populations,between ILs from different selection schemes,and G × T interaction using SAS PROC GLM [19].Student t-tests were performed to compare differences between the selected ILs and the recipient HHZ for measured traits under each water treatment.Selection efficiency was assessed for each selection scheme based on the number of ILs showing significantly improved trait values.

3.Results

3.1.Efficiencies of direct selection for improving ST and DT

The first round selection based on survival of individual plants for seedling ST in the screen-house resulted in 57(11.9%),49 (10.2%) and 56 (11.7%) plants from the HHZ/IR64,HHZ/AT354 and HHZ/C418 BC2F2populations,respectively(Fig.1).Progeny testing of the selected lines in the phytotron confirmed that the 25,28 and 29 lines from the three populations showed significantly improved ST as indicated by their significantly longer DSS and lower SST.This suggests a realized heritability of 0.439,0.571 and 0.518 from the single plant selection for seedling ST from the three BC2F2populations.

The initial screen for DT under the severe field drought conditions in Hainan resulted in 19 (4.0%),29 (6.0%) and 33(6.9%)plants with obviously higher fertility and GY than HHZ selected from the HHZ/IR64,HHZ/AT354 and HHZ/C418 BC2F2populations (Fig.1).However,the severe drought in the progeny testing under the controlled conditions of the greenhouse in Beijing killed HHZ (no yield),but 12,23 and 8 BC2F3lines from the three populations survived and produced seeds,resulting in a realized heritability of 0.632,0.793 and 0.242 from the single plant selection for DT from the three BC2F2populations in Hainan.When evaluated under the mild drought stress in Hainan during the 2011–2012 DS,8 of the 43 DT selected ILs showed significantly higher GY than HHZ,and none of them had lower GY than HHZ(Table 1),indicating that the selection for DT was highly effective.

Table 1-Numbers of HHZ introgression lines (ILs) selected for ST,HY and DT from three BC2F2 populations deviating significantly from the recipient HHZ for eight measured traits evaluated under drought stress(S)conditions of Hainan in the 2010-2011 dry season.

3.2.Indirect responses in nontarget traits of first round selection for HY,ST and DT

When the 189 ILs were evaluated under drought stress and normal irrigated conditions of Hainan during the 2011–2012 DS,water treatments (T) had highly significant effect on all measured traits,but this variation component varied considerably among different traits with R2ranging from 2.3%for PN to 45.7% for FNP.On average,the yield reduction caused by the drought stress was 20%for HHZ(the recipient)but 36.1%for the 189 ILs.Differences among different ILs (G) were highly significant for all measured traits and accounted for an average 36.6%of the total trait variation,ranging from 26.7%for FNP to 53.9% for PH.The T × G interaction was insignificant for all measured traits,indicating that all ILs performed consistently under drought stress and well watered conditions for the measured traits in this experiment.ANOVA also indicated that ILs from different populations showed significant differences for all measured traits except for PH,ranging from 2.3%for PN to 19.0%for GW.Similarly,different selection schemes had highly significant effects on the mean performances of the ILs for all traits except for SF and GY,ranging from 1.8%for PN to 38.4%for HD.Although all were highly significant,ILs selected from different populations (P) showed much greater trait variation than ILs obtained from different selection schemes (S).The P × S interaction was also significant for all measured traits,indicating that selection efficiency on any specific trait varied depending on the population(donor).

Table 2-Mean performances of 189 HHZ introgression lines (ILs) selected for three different traits from three BC2F2 populations for eight yield related traits evaluated in replicated experiments under normal irrigated(N)and drought stress(S) conditions in Hainan during the 2010-2011 dry season.

Table 3-Numbers of HHZ introgression lines (ILs) selected for ST,HY and DT from three BC2F2 populations that deviated significantly from the recipient HHZ for eight measured traits evaluated under the normal irrigated conditions of Hainan in the 2010-2011 dry season.

3.2.1.Indirect responses of selection for HY

Under normal irrigated conditions in Hainan,the 64 ILs selected for HY in Beijing had an average yield of 24.9 g per plant,or 13.2%higher than HHZ(Table 2).Of these,8 ILs had significantly higher GY than HHZ,resulting primarily from increased SNP/FNP (Table 3).The remaining ILs had the same GY as HHZ.Interestingly,a significant portion of these ILs showed early heading and reduced GW as indirect responses to selection for HY.Under water stress,the 64 ILs had a mean GY of 14.8 g per plant,or 15.9% lower than HHZ.The numbers of ILs that had significantly higher and lower yields than HHZ were 8 and 10(Table 1) with most DT ILs coming from the HHZ/C418 population and most drought sensitive ILs coming from HHZ/AT354.Many lines in this group of ILs showed early heading,reduced height and reduced fertility under stress(Table 1).

3.2.2.Indirect responses of selection for ST

Under normal irrigated conditions in Hainan,the 82 ST selected ILs had an average GY of 24.7 g per plant,or 12.1%higher than HHZ (Table 2).Of these,10 ILs had significantly higher GY than HHZ,resulting primarily from increased SNP/FNP,PN and PH(Table 3).Only two ILs had significantly lower GY than HHZ.Again,many of these ILs showed early heading,reduced GW and lower fertility as indirect responses to selection for ST.Under water stress,the 82 ILs had a mean GY of 16.0 g per plant,or 9.1% lower than HHZ.The numbers of ILs that had significantly higher and lower GY than HHZ were 14 and 18 (Table 1),which were roughly equal from the three populations.Many of these ST ILs showed early heading and reduced SF/FNP under stress(Table 1).

3.2.3.Indirect responses of selection for DT

This group of 43 ILs had gone through two rounds of selection for DT,one in Hainan and one in Beijing.Under the severe drought of Beijing that killed HHZ (100% yield reduction),the 43 ILs had an average GY of 9.0 g per plant,or a reduction of 70.2% compared with their GY in the irrigated control(Table 4).Under normal irrigated conditions,the 43 ILs had an average GY of 25.4 g,or 9.9% higher than HHZ.Of these,only eight ILs had significantly higher average GY than HHZ and the remaining ILs had the same GY as HHZ(Table 3).

In Hainan,the 43 ILs had an average GY of 24.0 g per plant,or 9.1% higher than HHZ under irrigated conditions (Table 2).Of these,five ILs had significantly higher GY than HHZ,resulting primarily from increased SNP and PH (Table 3).The remaining ILs had the same GY as HHZ.Again,early heading was an indirect response to selection for DT in 20 of the 43 ILs(Table 3).Under water stress,the 43 ILs had a mean GY of 16.2 g per plant,or 8.0% lower than HHZ.Eight ILs had significantly higher GY than HHZ,most of which were from population HHZ/C418 (Table 1).None of these ILs had lower GY than HHZ and 15 ILs showed delayed heading.

ANOVA of the combined data from Beijing and Hainan indicated that the differences among the ILs (G) were highly significant for all measured traits and explained,on average,17.0% of the total phenotypic variation,ranging from 8.5% for PN to 31.5% for HD.The difference among locations (L) was highly significant for all traits and explained an average of 14.0%of the total variation,ranging from 1.9%for SF to 36.9%for PN.The difference between the two water treatments(T)was highly significant for all traits and accounted for an average 32.6% of the total variation,ranging from 3.9%for PN to 56.0%for GY.The G × T,G × L,L × T and G × T × L interactions were also highly significant for all traits except HD and PN (G × T × L),and explained an average of 7.0%,7.4%,4.3% and 6.1% of the total trait variation,respectively.

Table 4-Mean performances of 43 HHZ introgression lines(ILs)selected for drought tolerance from three BC2F2 populations for yield related traits evaluated in replicated experiments for yield traits under normal irrigated(N)and drought stress(S)conditions of Beijing in the 2010 summer(Mean ± SD).

3.3.Promising ILs developed

Table 5 shows the mean trait performances of 16 promising HHZ ILs that had significantly higher GY and/or better DT than HHZ in at least one location.These included 10 DT selected ILs,3 ST selected ILs for and 3 HY selected ILs,respectively.Of these,WT185 was the best and was originally selected for DT but showed significantly higher GY than HHZ under drought and non-stress conditions in both Hainan and Beijing.

4.Discussion

HHZ is a high yielding and widely adapted variety currently grown on 3,500,000 ha in southern and central China.It also performs well in many countries in tropical Asia and Africa(data not shown).However,it does not have good tolerance to many abiotic stresses.This study reports part of our efforts to convert it into a green super rice (GSR) variety with tolerance to multiple abiotic stresses using a BC breeding strategy.Consistent with previous results [14–16],the development of

many HHZ ILs with significantly improved DT,ST or HY demonstrated that BC breeding and phenotypic selection were effective for improving single complex traits in rice.Furthermore,direct comparison between the ILs and HHZ for yield performance and related traits under drought stress and non-stress conditions across different environments led us to several important conclusions regarding how to improve selection efficiency and overall genetic gain when aiming to improving multiple complex traits in a BC breeding program.

Table 5-Mean trait performances of 16 promising ILs selected from three BC populations with significantly higher yields than HHZ in at least two environments under normal irrigated and drought stress conditions in Hainan and Beijing.

Firstly,our results indicated that the primary target traits should be selected first in the target environments.This was reflected by the huge differences between ILs generated from the three selection schemes(Table 1)and by the fact that the most promising HHZ ILs showing significantly improved DT and yield in Hainan were originally DT selected(Table 5).This was not surprising since the initial selection for DT was carried out in Hainan,whereas the yield performances of the ST and HY selected HHZ ILs under drought and non-stress conditions in Hainan were indirect responses.Interestingly,we observed positive gains of 12.2% and 12.5% in GY under normal conditions in Hainan as indirect responses to selection for ST and HY in Beijing,and found no evidence for a yield penalty associated with DT in the tested HHZ ILs (Table 3).

Secondly,our results indicated that selection for DT in the DS in Hainan practiced in many Chinese rice breeding programs should be largely effective.In this study,the overall level of G × E interaction accounted for only(14.2%)of GY in the 43 DT selected ILs,3.4%,6.1%and 4.7%of which was attributed to the G × T,G × L and G × T × L interactions.This low level of G × E interaction for DT was at least partially attributable to the greater genetic uniformity of the BC progeny because the large G × E interaction commonly observed in breeding for DT in rice results primarily from the complexity of the drought environments[20]and at least partially from the complex genetic and physiological mechanisms of DT in rice[21].

Thirdly,we did not observe a strong negative association between DT with GY under normal conditions as all DT selected ILs had the same or higher GY than HHZ under the normal irrigated conditions in Beijing or Hainan (Tables 1 and 3).However,we noted that 15 (～35%) of the DT selected lines showed delayed heading under drought in Hainan,whereas most(78.1%)ST and HY selected ILs showed significantly earlier heading(Table 1).Curiously,increased plant height was observed as an indirect response to selection for DT and cold tolerance(CT)in the japonica backgrounds[16,22],but was not observed in this study.Interestingly,under normal irrigated conditions in Hainan,20(46.5%)DT selected ILs,16(19.5%)ST selected ILs and 20(31.3%)HY selected ILs showed earlier heading.All DT selected ILs showed earlier heading under normal irrigated conditions in Beijing (Table 3).This suggests that the donors contributed different genetic and physiological mechanisms for DT in HHZ(indica)than those for DT in japonica backgrounds[16,22].

Fourthly,our results indicated that parental selection is critically important for the success of a BC breeding program.While widely adaptable superior commercial lines should be used as recurrent parents,the choice of donors of target traits may be more difficult.In this study,the japonica donor,C418 was apparently a better donor than the two tropical indica donors(IR64 and AT354)in contributing promising DT and HY progeny in Hainan.This was surprising since none of the donors was superior for the target traits.In two separate experiments,we found that indica donors tend to contribute more trait enhancing alleles for DT and CT than japonica lines [16,22].Thus,exploiting the genetic diversity in the subspecific gene pools using BC breeding will be of great importance for future genetic improvement of complex traits in rice.

Finally,the presence of significant amounts of useful genetic variation for yield related traits under drought and non-stress conditions among ILs within the same or different BC populations indicates that considerable genetic gain can be achieved through selection for secondary target traits among the ILs.However,initial selection for different traits resulted in ILs that varied considerably for the measured traits,suggesting that selection efficiency for secondary target traits would be very different for ILs selected for different primary traits.Selection for secondary target traits can be done more effectively by screening resistances/tolerances to different biotic and abiotic stresses and quality traits through replicated progeny testing of the ILs.In this regard,the developed ILs provide useful materials for genetic and molecular dissection of complex traits using functional genomic tools [23] and for developing high yielding GSR cultivars with multiple green traits using new and innovative molecular breeding strategies[24].

5.Conclusions

It was demonstrated that BC breeding and phenotypic selection were effective for simultaneous improvement of multiple complex traits(HY,DT and ST)in rice.The primary target traits should be selected first in the target environments (TEs) to achieve the maximum genetic gain.BC breeding for DT in rice was almost equally effective by strong phenotypic selection in the TEs and in the winter-season nursery in Hainan.Considerable genetic gain can be achieved by selection for secondary target traits among the ILs with the primary traits.Exploiting genetic diversity in the subspecific gene pools will be of great importance for future genetic improvement of complex traits in rice.Finally,the ILs developed in this study provide useful materials for future genetic/genomic dissection and molecular breeding for genetic complex traits.

This work was funded by the National High Technology Research and Development Program of China (2012AA101101)from the Ministry of Science and Technology of China,the National Science Foundation Project(30570996),the Program of Introducing International Super Agricultural Science and Technology (#2011-G2B) from the Ministry of Agriculture of China,and the Bill&Melinda Gates Foundation Project(OPP51587).

[1] P.Lucca,R.Hurrell,I.Potrykus,Fighting iron deficiency anemia with iron-rich rice,J Am Coll Nutr 21(2002) 184–190.

[2] A.Munir,M.Hanjra,Q.Ejaz,Global water crisis and future food security in an era of climate change,Food Policy 35(2010) 365–377.

[3] B.A.M.Bouman,T.P.Tuong,Field water management to save water and increase its productivity in irrigated lowland rice,Agric Water Manag 49(2001) 11–30.

[4] Z.G.Bai,D.L.Dent,L.Olsson,M.E.Schaepman,Proxy global assessment of land degradation,Soil Use Manag 24(2008)223–234.

[5] E.A.Bray,J.Bailey-Serres,E.Weretilnyk,Responses to abiotic stresses,in:W.Gruissem,B.Buchannan,R.Jones(Eds.),Biochemistry and Molecular Biology of Plants,American Society of Plant Physiologists,Rockville,MD,USA,2000,pp.1158–1203.

[6] P.Q.Craufurd,J.M.Peacock,Effect of heat and drought stress on sorghum,Exp Agric 29(1993) 77–86.

[7] Y.Jiang,B.Huang,Drought and heat stress injury to two cool season turfgrasses in relation to antioxidant metabolism and lipid peroxidation,Crop Sci 41(2001) 436–442.

[8] A.S.Moffat,Finding new ways to protect drought-stricken plants,Science 296 (2002) 1226–1229.

[9] M.Pitman,A.L?uchli,Global impact of salinity and agricultural ecosystems,in:A.Lauchli,U.Luttge(Eds.),Salinity:Environment–Plants–Molecules,Kluwer Academic Press,Netherlands,2002,pp.4–20.

[10] G.J.Pan,S.Q.Chen,C.Y.Song,G.L.Wang,B.Pan,Study on relationship between resistance to blast and yield traits in early japonica rice in cold region,Chin Agric Sci Bull 25(2009)236–238.

[11] Y.L.Xia,M.F.Lai,Y.Zeng,Y.Qi,Correlation of resistance to late leafspot and yield characters in groundnut,J Peanut Sci 33 (2004) 28–31,(in Chinese with English abstract).

[12] G.H.Yin,Y.J.Shen,Z.J.Kang,F.S.Zheng,Z.X.Liu,Screening on varieties of spring maize in the semiarid region of west Liaoning,Chin Agric Sci Bull 27(2011) 195–198.

[13] Z.Zhou,H.Li,Y.Sun,D.Q.Huang,L.H.Zhu,D.C.Lu,K.H.Li,J.L.Xu,S.C.Zhou,Z.K.Li,Effect of selection for high yield,drought and salinity tolerances on yield-related traits in rice(Oryza sativa L.),Acta Agron Sin 36(2010) 1725–1735,(in Chinese with English abstract).

[14] A.J.Ali,J.L.Xu,A.M.Ismail,B.Y.Fu,C.H.M.Vijaykumar,Y.M.Gao,J.Domingo,R.Maghirang,S.B.Yu,G.Gregorio,S.Ynaaghihara,M.Cohen,B.Carmen,D.Maekill,Z.K.Li,Hidden diversity for abiotic stress tolerances in the primary gene pool of rice revealed by a large backcross breeding program,Field Crops Res 97(2006)66–76.

[15] H.R.Lafitte,Z.K.Li,C.H.M.Vijayakumar,Y.M.Gao,Y.Shi,J.L.Xu,B.Y.Fu,S.B.Yu,A.J.Ali,J.Domingo,R.Maghirang,R.Torres,D.Mackill,Improvement of rice drought tolerance through backcross breeding: evaluation of donors and selection in drought nurseries,Field Crops Res 97 (2006)77–86.

[16] Y.X.He,T.Q.Zheng,X.B.Hao,L.F.Wang,Y.M.Gao,Z.T.Hua,H.Q.Zhai,J.L.Xu,Z.J.Xu,L.H.Zhu,Z.K.Li,Yield performances of japonica introgression lines selected for drought tolerance in a BC breeding programme,Plant Breed 129 (2010) 167–175.

[17] S.Yoshida,D.A.Forno,J.H.Cock,K.A.Gomez,Laboratory Manual for Physiological Studies of Rice,3rd edn IRRI,Manila,Philippines,1976.

[18] International Rice Research Institute(IRRI),Standard Evaluation System for Rice,IRRI,Manila,Philippines,1996.

[19] SAS Institute,SAS/STAT User's Guide,SAS Institute,Cary,North Carolina,1996.

[20] S.Fukai,M.Cooper,Development of drought-resistant cultivars using physiomorphological traits in rice,Field Crops Res (1995) 67–86.

[21] Z.K.Li,J.L.Xu,Breeding for drought and salt tolerant rice(Oryza sativa L.): progress and perspectives,in:M.A.Jenks,P.M.Hasegawa,S.M.Jain (Eds.),Advances in Molecular Breeding toward Drought and Salt Tolerant Crops,Springer Press,Netherlands,2007,pp.531–564.

[22] L.J.Meng,X.Y.Lin,J.M.Wang,K.Chen,Y.R.Cui,J.L.Xu,Z.K.Li,Simultaneous improvement of cold tolerance and yield of temperate japonica rice(Oryza sativa L.) by introgression breeding,Plant Breed,2013.(accepted for publication).

[23] Z.K.Li,B.Y.Fu,Y.M.Gao,J.L.Xu,J.Ali,H.R.Lafitte,Y.Z.Jiang,J.D.Rey,C.H.M.Vijayakumar,R.Maghirang,T.Q.Zheng,L.H.Zhu,Genome-wide introgression lines and a forward genetics strategy for functional genomic research of complex phenotypes in rice,Plant Mol Biol 59(2005)33–52.

[24] Z.K.Li,Strategies for molecular rice breeding in China,Mol Plant Breed 3(2005)603–608,(in Chinese with English abstract).