Crop Science Illumina
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 30 July 2007
Published in Crop Sci 47:1561-1573 (2007)
© 2007 Crop Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Rane, J.
Right arrow Articles by Joshi, A. K.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Rane, J.
Right arrow Articles by Joshi, A. K.
Agricola
Right arrow Articles by Rane, J.
Right arrow Articles by Joshi, A. K.
Related Collections
Right arrow Field evaluation techniques
Right arrow Plant and Environment Interactions
Right arrow Crop Ecology

CROP ECOLOGY, MANAGEMENT & QUALITY

Performance of Yield and Stability of Advanced Wheat Genotypes under Heat Stress Environments of the Indo-Gangetic Plains

Jagadish Ranea,*, Raj Kumar Pannub, Virinder Singh Sohuc, Ran Singh Sainid, Banwari Mishraa, Jag Shorana, Jose Crossae, Mateo Vargase and Arun Kumar Joshif

a Directorate of Wheat Research, Karnal, Haryana, India
b CCS Haryana Agriculture Univ., Hisar, Haryana, India
c Punjab Agriculture Univ., Ludhiana
d A.R.S., Rajasthan Agriculture Univ. Durgapura, Jaipur, Rajasthan, India
e Biometrics and Statistics Unit, Crop Informatics Lab., International Maize and Wheat Improvement Center (CIMMYT) Apdo. Postal 6-641, C.P. 06600, D.F. Mexico
f Dep. of Genetics and Plant Breeding, Institute of Agricultural Sciences, Banaras Hindu Univ., Varanasi 221005, India

* Corresponding author (jagrane{at}hotmail.com).


   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
A set of 25 advanced breeding lines and released varieties of wheat (Triticum aestivum L.) developed by different breeding centers in India were assessed for their adaptation in 18 different environments across the Indo-Gangetic plains. The study was aimed at identifying genotype(s) with high yield stability across the environments in general and heat stress environments in particular. Jaipur and Varanasi were hotter than any other locations considered in this study. Considerable intralocation variation in genotypic response pattern was observed over the years and dates of sowing, and this was more conspicuous at Varanasi. Longer crop duration and short grain growth duration at Varanasi were in contrast to shorter crop duration and relatively longer grain growth period that supported better grain growth at Jaipur. The genotype x environment interaction biplots for grain yield revealed that genotypes Raj 3765 and Raj 4027, developed at Jaipur, were more stable across all environments. This was due to their adaptability to high-temperature environments, and hence they are being proposed as promising germplasm sources for late-sown and/or warmer environments. Since the pattern of genotypic response observed at Jaipur was not similar to that observed at Varanasi, it is suggested that a common breeding strategy, if any, should emphasize grain yield stability for breeding for high-temperature tolerance. This can also take care of intralocation variation in genotypic response over the years and dates of sowing at Varanasi.

Abbreviations: CIMMYT, International Center for Maize and Wheat Research • GxE, genotype x environment • PLS, partial least squares • SRG, sites regression model


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
WHEAT (Triticum aestivum L.) is the second most important crop of India, and improvement in its productivity has played a key role in making the country self-sufficient in food production. However, in the past decade there has been marginal increase in the productivity of wheat, particularly under environments relatively favorable for growth and development of wheat (Nagarajan, 2005; Joshi et al., 2007). On the other hand, there is substantial scope for improvement in productivity under unfavorable environments that are characterized by a significant presence of abiotic stresses such as high temperature (Aggarwal, 1991; Joshi et al., 2007).

Continuous high-temperature stress for wheat has been defined as when the mean average temperature of the coolest month is greater than 17.5°C (Fischer and Byerlee, 1991). Terminal heat stress largely refers to rise in temperatures at the time of grain growth. In wheat, high temperatures (>30°C) after anthesis can decrease the rate of grain-filling (Al-Khatib and Paulsen, 1984; Randall and Moss, 1990; Stone et al., 1995; Wardlaw and Moncur, 1995), while high temperatures imposed before anthesis can also decrease yield (Wardlaw et al., 1989a; Hunt et al., 1991). Under controlled experiments, grain yield of wheat per spike was reduced by 3 to 4% per 1°C increase in temperature over 15°C. (Wardlaw et al., 1989a,b). The effect of short periods of exposure to high temperatures (>30°C) on wheat grain yield are thought to be equivalent to a 2 to 3°C warming in the seasonal mean temperature (Wheeler et al., 1996). Also up to a 23% reduction in grain yield has been reported from as little as 4 d of exposure to very high temperatures (Randall and Moss, 1990; Stone and Nicolas, 1994). Periods at such high temperatures occur frequently during grain-filling in both Mediterranean and continental climates. The heat-prone mega-environment—classified as ME5 by the International Maize and Wheat Improvement Center (CIMMYT), based on agroecological parameters (Braun et al., 1992), shows a range of heat profiles when comparing sites in different countries.

According to an estimate, there are currently around 9 million ha of wheat in tropical or subtropical areas (Lillemo et al., 2005) that experience yield losses due to high-temperature stress. The areas are in countries including Bangladesh, India, Nigeria, Uganda, Sudan, and Egypt that have long traditions of cultivating wheat. However, the current estimates indicate that in India alone around 13.5 million ha of area is heat stressed (Joshi et al., 2007). In India, incidences of high temperatures at the time of grain-filling are more pronounced when sowing of wheat is delayed due to delay in harvest of highly remunerative preceding crops such as scented rice (Tandon 1994; Nagarajan and Rane 1997; Sharma et al., 1997; Rane et al., 2000; Joshi et al., 2002). Intensity of high temperatures is likely to become much larger if current trends and future predictions about global warming continue (Kattenberg et al., 1995). Furthermore, current recommendations for crop management practices that can reduce heat stress to plants rely heavily on additional inputs, especially irrigation water (Badaruddin et al., 1999). This is an additional concern given that water resources around the globe are shrinking (World Meteorological Organization, 1997), and there is need of more sustainable and environmentally friendly approaches for increasing productivity. Hence, developing genotypes that possess thermal tolerance is one of the major priorities of wheat improvement programs of the Indo-Gangetic Plains.

Genotypes suited to late-sown environments having heat stress have been developed primarily through empirical selection with greater emphasis on grain yield and biotic stresses. Genotypes that perform better in highly stressed environments at one location may perform better at similar locations elsewhere (Reynolds et al., 1994). Superior genotypes developed at particular locations of India have been found to be equally good or superior to locally developed genotypes at other national centers, indicating wide adaptability (Nagarajan, 2001). This can be attributed to their yield potential under favorable environments or tolerance to abiotic stresses at unfavorable environments. However, high-yielding genotypes do not perform on par with abiotic stress–adapted genotypes when yield is depressed below a crossover point (Blum, 1996). Although, approaches other than that based on breeding for yield per se have been proposed (Reynolds et al., 1998), yield and yield traits continue to be important to measure the success of a genotype in heat-stressed environments. A genotype with stable and high yield across the environments would be more suitable as a cultivar and also as a donor parent for further breeding in hot environments that vary over the years and within a particular year across locations.

Interpretation of performance of a number of genotypes in a broad range of environments is always affected by large genotype x environment (GxE) interactions (Gauch and Zobel, 1996). The ordinary ANOVA describes only the main effects effectively and can test the significance of the GxE interaction but provides no insight into the particular patterns of genotypes or environments that give rise to the interaction. Multiplicative models for multienvironment trials have been used for studying GxE interaction, examining genotypic yield stability and adaptation and for developing methods for clustering sites or cultivars into groups with statistically negligible crossover GxE interaction (Crossa and Cornelius, 1997; Crossa et al., 2002). Multiplicative models have an additive (linear) component (i.e., intercept, main effects of sites and/or genotypes) and a multiplicative (bilinear) component (GxE) and thus are also named linear-bilinear models. A type of linear-bilinear model suitable for grouping sites and cultivars without cultivar rank change is the sites regression model (SRG). Biplots obtained from graphing the first two components of the multiplicative part of the SRG are useful for summarizing and approximating patterns of response that exist in the original data (Gabriel, 1971). Recently, Crossa et al. (2002) showed that for SRGs, the biplot of the first two multiplicative components graphs the interaction variation due to noncrossover GxE interaction versus the interaction variation due to crossover GxE interaction explainable by a second bilinear term. They further demonstrated how the SRG biplots could be used for identifying subsets of sites and genotypes with noncrossover GxE interaction. Results of international yield trials by CIMMYT grown over 40 environments confirmed that the main factor determining GxE interaction in hot climates was relative humidity (Reynolds et al., 1998), while more subtle effects were identified explaining GxE interaction within the hot, low relative humidity environment (Vargas et al., 1998). Subsequently, analysis of CIMMYT's High-Temperature Wheat Yield Trial, which consisted of 233 individual yield trials grown at 101 locations, showed that the main genotype clusters corresponded to three main types of environment, temperate, continuous heat stress, and terminal heat stress, while relative humidity was also an important factor determining GxE interaction within some of these clusters (Lillemo et al., 2005).

Complexities associated with environmental variables such as radiation, high temperature, and relative humidity, often make it difficult to understand GxE interactions. Such problems were addressed through partial least squares (PLS) regression while explaining physiological factors associated with GxE interaction in wheat grown across different locations (Reynolds et al., 2002).

This study was conducted to investigate a set of 25 advanced breeding lines and released varieties developed by different breeding centers in India to identify genotypes with high yield stability under high-temperature environments that occur due to delay in planting.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Experimental Data
A total of 25 advanced breeding lines (Table 1) and cultivars of wheat developed at different research stations across India (Table 2) were used in this study (Table 1). Some of the genotypes selected from CIMMYTs nurseries and trials were also included. Many of these genotypes were developed for late-sown environments where high temperatures (30–35°C) occur during grain-filling in wheat. Field trials were conducted at five locations: Karnal, Ludhiana, Hisar, Jaipur, and Varanasi. Details of latitude, longitude, altitude, and soil type for these locations are given in Table 2. Trials were conducted for a period of two consecutive years in two dates of sowing (normal and late) commencing in the 2000–2001 cropping season at all the locations except at Jaipur, where trials were conducted only during the first year. Dates of planting and temperature regimes at different locations are given in Table 3. Normal sowing of wheat in the Indo-Gangetic plains of India is considered to range from 5 to 25 November; whereas very late sowing is considered to be beyond 15 December to the end of December. Wheat sowing beyond December is generally not recommended. Following this approach, a total of 18 environments were obtained by combinations of year, location, and time of sowing.


View this table:
[in this window]
[in a new window]

  		Table 1. Origin and pedigree of genotypes.

 

View this table:
[in this window]
[in a new window]

  		Table 2. Soil classification, latitude, longitude, and altitude of five locations used in evaluation of heat tolerance in 25 wheat genotypes.

 

View this table:
[in this window]
[in a new window]

  		Table 3. Dates of sowings and mean values for daily minimum and maximum temperatures recorded during pre- and postanthesis stages of crop growth.

 
The experimental design in each environment was a 5 x 5 square lattice design with two replications. Each plot consisted of six rows with 23 cm between the rows. Length of the plot was 3 m at all the locations. Each plot was sown at a seeding rate of 100 kg ha–1. Nitrogen, phosphorous, and potash were applied at a rate of 60 kg ha–1 each at seeding. An additional 60 kg ha–1 nitrogen was applied at the time of tillering. At all the locations, irrigation was provided to the crop whenever necessary to avoid moisture stress.

Observations were recorded for phenology and yield traits for all the genotypes at all locations. The traits were date of anthesis (d), date of maturity (d), plant height (cm), number of productive tillers m–2, biomass (g m–2), grain yield (g m–2), thousand grain weight (g), number of grains per spike, grain weight per spike (g), and grain growth duration (d to physiological maturity – d to anthesis). Anthesis dates were recorded for each cultivar and defined as the date when 75% of the wheat spikes had visible anthers. Physiological maturity was recorded when rachis of 75% of the spikes in each plot lost green color. Grain yield and biomass from each plot were obtained when plants were fully dry and grains were at 12% moisture. Five main shoot spikes were sampled from each plot, and grains were separated by hand for determining single grain weight, grain number per spike, and grain weight. Grains were dried in an oven at 65°C for 72 h before determining grain weight.

Daily mean maximum and minimum temperatures were considered for characterizing environments. Mean values of temperatures for the period corresponding to the preanthesis phase were calculated by taking into consideration the date of planting and the earliest day on which anthesis was recorded, while the same for the postanthesis period were calculated on the basis of date of earliest anthesis and the day when the last physiological maturity was recorded for each environment.

Statistical Analyses
Assessment of Thermal Stress in Different Environments
To differentiate environments on the basis of magnitude of temperature stress, five temperature parameters, listed in Table 3, were used for cluster analysis. The classification was based on the Ward (1963) method, and dissimilarities between sites were measured by squared Euclidean distance. The cutoff point for grouping was decided on the basis of differences in temperature variables within and between the groups as revealed by one-way analysis of variance and post hoc tests conducted by using SPSS software (SPSS, 2002). The mean maximum temperatures in the seventh week after planting were considered, as they could substantially explain variation in grain yield across the environments as revealed by correlation studies.

SRG Biplot
Stability analysis using the SRG was performed for the response of the 25 genotypes in the 18 environments that resulted from all the possible combinations of the sites, the years of testing, and the dates of sowing. Environments were denoted by the first letter of the site followed by a number indicating the first or second year and suffixed by a letter indicating time of planting, that is, T for the first or normal and L for the second or late planting. For example, J1T represents data at Jaipur in the first year under the first or normal planting.

The SRG is

Formula
where, yij is the mean of the ith cultivar in the jth environment for g cultivars and e sites; i = 1, 2, ...; g and j = 1, 2, ... e); µj is the site mean; {lambda}k ({lambda}1 ≥ {lambda}2 ≥ ... ≥ {lambda}t) are scaling constants (singular values) that allow the imposition of orthonormality constraints on the singular vectors for cultivars {alpha}k = ({alpha}1k,... {alpha}gk)' and sites {gamma}k = ({gamma}1k,...,{gamma}ek)', such that {sum}i{alpha}ik2={sum}j{gamma}jk2=1 and {sum}iáikáik'={sum}j{gamma}jk{gamma}jk'=0 for k != k'; {alpha}ik and {gamma}jk, for k = 1, 2, 3, ... are called "primary," "secondary," "tertiary," and so on effects of the ith cultivar and the jth site, respectively; Formulaij is the residual error assumed to be normally and independently distributed with 0 means and variance {sigma}2/r (where {sigma}2 is the pooled error variance and r is the number of replicates). The number of bilinear terms is t ≤ min (g, e). Estimates of the multiplicative parameters in the kth bilinear term are obtained as the kth component of the deviations from the additive part of the model. In the SRG, only the main effects of cultivars plus the GxE interaction are absorbed into the bilinear terms.

PLS and the Biplot
The PLS extracts the main variation patterns from one data table X that have relevance also to another data table Y from the same individuals. The PLS can be seen as an extension of principal component analysis because it allows extraction of the main variation patterns within X and Y, permitting study of the structure between X and Y. Thus, the latent variables extracted are the essence contained in the variables of X that are also relevant to Y, so that both matrices are modeled simultaneously. Thus the bilinear representation of X and Y are as follows:

Formula

Formula
where the matrix T contains the Z scores, matrix P contains the Z loadings, matrix q contains the Y loadings, and F and E are the residual matrices. The basic idea is that the relationship between Y and X is transmitted through the latent variables t. The choice of what is X and what is Y does not have to follow the traditional X = cause and Y = effect, like in the classical regression where X is called the independent (or regressor) variable and Y is called the dependent variable (Martens and Martens, 2001). In the context of multilocation trials, the Y matrix consists of variables (e.g., grain yield, days to maturity) measured on genotypes in different environments, and the X matrix comprises covariables that have been measured in either the genotypes (e.g., days to anthesis, physiological maturity) or in the environments (e.g., minimum temperatures, maximum temperatures). In this context, the covariables measured in X can explain some of the variability existing in Y. In other words, genotypic or environmental covariables can help us to explain GxE interaction.

The results of the PLS can be graphically displayed in the form of biplots (Gabriel, 1971) in which coordinates for environments, genotypes, and environmental covariables corresponding to the first two PLS components are simultaneously depicted by vectors in a space with starting points at the origin (0,0) and end points determined by the value of the coordinate. Genotypes and environments having the same directions have positive interactions, and those having opposite directions have negative interaction. Details of the univariate and multivariate PLS algorithms are given by Vargas et al. (1998, 1999).


   RESULTS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Environments
The cluster analysis by Ward's method could classify 18 environments into three distinct groups on the basis of five temperature variables considered in this study (Fig. 1 ). Environment Group I comprised all the timely sown environments except L1T and H2T of Karnal, Ludhiana, and Hisar. L1T and H2T were as hot as late-sown environments that were grouped in Environment Group II, contradicting the presumption that all the crop environments in the second date of sowing always exposed genotypes to hotter environments than the first one. This was also evident from the composition of Environment Group III, which had both the normal and late-sown environments of Jaipur and Varanasi. The Environment Group III was significantly hotter than the other two environments throughout the crop duration, as revealed by ANOVA (Fig. 1). However, the difference between the first and second environment was not significant for any of the temperature variables except for the maximum temperatures in the seventh week after planting and also after anthesis. The mean maximum after anthesis for Environment Group III was 35.9°C, while it was 29.0°C for Environment Group I. A similar trend was observed for mean minimum temperatures, but the difference between the seventh week mean maximum temperatures recorded in Environment Group III and the same in the Environment Group I was as high as 9°C, which indicates that the crop was exposed to a significantly high level of thermal stress even at early stages of growth in the former group of environments.


Figure 1
View larger version (24K):
[in this window]
[in a new window]

  		Figure 1. Dendrogram of 18 environments based on five temperature parameters (seventh week after planting [7WM], mean minimum temperature before anthesis [BAMN], mean maximum temperature before anthesis [BAMX], mean minimum temperature after anthesis [AAMN], and mean maximum temperature after anthesis [AAMX]) analyzed by Ward's incremental sum of squares method. The table superimposed on the graph contains mean values of temperature variables for Environment Group I (EnvGr-I), Environment Group II (EnvGr-II), and Environment Group III (EnvGr-III). Letters used in superscripts (a, b, c) indicate significance of difference in mean values at p > 0.05 as revealed by post hoc test.

 
GxE as Revealed by SRG
A combined analysis of variance of all the traits of 25 genotypes in different environments (Table 4) indicated that substantial variation existed for environment, environment x genotype, environment x sowing, and environment x genotype x sowing for yield, yield components, and phenology of genotypes.


View this table:
[in this window]
[in a new window]

  		Table 4. Analysis of variance for 11 traits of 25 wheat genotypes across 18 environments that resulted from combination of sites and dates of planting over 2 yr at four locations and for 1 yr at one location.

 
The grain yield (g m–2) of 25 genotypes for each environment (Table 5) varied from 120.0 (WH 730 in J1T2) to 665.0 (HW2044 in L1T1) with an overall mean of 418.3. Three genotypes, 2 (PBW 497), 17 (Raj 3765), and 10 (HW 2044), had the highest yield under late-sown conditions with a low to moderate decline in performance due to late sowing (Table 6). However, the lowest decline due to late sowing across environments was displayed by two other genotypes, 25 (WR 704) and 5 (HD 2285), which had relatively low grain yield.


View this table:
[in this window]
[in a new window]

  		Table 5. Mean performance of grain yield of 25 genotypes in 18 environments (over two years and two dates of sowing at four locations and for one year at one location).

 

View this table:
[in this window]
[in a new window]

  		Table 6. Mean yield and tillers m–2 under timely, late-sown, and overall basis and their percent decline due to delay in sowing of 25 genotypes at five locations.

 
The SRG analysis of grain yield across 18 environments demonstrated that environment L1T contributed the most to the GxE interaction of grain yield (longest arrow), followed by environments K1T, K2L, L1L, H1T, and V1T (Fig. 2 ). Wide angles between the vectors of V1T and all other environments except J1T in the biplot indicate a distinct response pattern of genotypes at Varanasi, while some degree of similarity between V1T had J1T is evident from narrow angles between their vectors in the SRG biplot. Interestingly, variation within the location is highly conspicuous at Varanasi, where environments in the first year were different from those in the second year.


Figure 2
View larger version (20K):
[in this window]
[in a new window]

  		Figure 2. Sites regression model biplot of GxE interaction of 25 wheat genotypes (1–25) across 18 environments shown in Table 2.

 
Considering that the first component of the SRG analysis accounted for the noncrossover GxE interaction variability and that the second components is due to the crossover GxE variance (Yan and Hunt, 2001), the ideal genotype is the one with the highest values for the first component and a value close to zero for the second component. The most stable genotypes, as revealed by this analysis, were 3 (CBW 12), 17 (Raj 3765), 22 (Raj 4027), 18 (Raj 4000), 9 (HUW 543), and 2 (CBW 09). Grain yield of Genotype 10 (HW 2004) was substantially high but was not stable across environments. The most unstable grain yield was observed in Genotypes 1 (7C Frango 60), 7 (HD 2428), 13 (NIAW 845), 20 (Raj 4014), 23 (Raj 4028), 24 (WH 730), and 25 (WR 704).

Temperatures and GxE Interaction
The first factor of PLS largely explained variance due to all four temperature variables used for analysis (Table 7). The mean maximum temperatures recorded after anthesis explained more variance than the same recorded before anthesis. Maximum temperatures recorded in the seventh week after anthesis also played a crucial role in determining grain yield. X loadings of all the temperature variables were highly positive (Table 8). The length of the vectors from origin of the biplot indicate that all the temperature variables significantly influenced GxE interaction. Influence of minimum temperature on GxE interaction appears to be independent of influence due to maximum temperature, as indicated by the wide angle between the vectors for these parameters in PLS biplot. Influence of maximum temperature on GxE interaction before anthesis was greater than that after anthesis. The narrow angle between the vectors for three variables of maximum temperatures indicates that their influences on GxE interaction were almost similar.


View this table:
[in this window]
[in a new window]

  		Table 7. Proportion of total X variances explained by the first and second factors and X loadings of genotypic variables obtain through partial least squares analysis of Y matrix for cultivar (25) vs. environment (18) grain yield, and X matrix for cultivar (25) vs. mean values for genotypic variables.

 

View this table:
[in this window]
[in a new window]

  		Table 8. Proportion of total X variances explained by the first and second factors and X loadings of environmental variables obtain through partial least squares analysis of Y matrix for cultivar (25) vs. environment (18) grain yield, and X matrix for cultivar (25) vs. mean values for temperature variables.

 
The response pattern of Genotype 4 (GW 173) was significantly influenced by minimum temperatures either before or after anthesis. Maximum temperatures influenced Genotype 10 (HW 2044), however, its yield tended to reduce whenever there was rise in minimum temperatures. A group of genotypes comprising 6(HD 2428), 7(HUW 234), 8(HUW 510), 9(HUW 543), and 23(WH 730) had significantly contrasting response to temperature regimes as compared to another group comprising 1(7cFrang 60), 3(CBW 12), 15(PBN 51), 19(Raj 4012), and 22(Raj 4027), as they were placed in quadrangles opposite to each other in the PLS biplot. High temperatures negatively influenced the second group of genotypes. The PLS biplot could effectively differentiate the hottest environments from the cool environments obtained by Ward's method. However, GxE interaction of hot environments in the second year of testing at Varanasi was influenced largely by high mean minimum temperatures, while the same in the rest of the hot environments was influenced by high mean maximum temperatures.

PLS Analysis of Genotypic Variables
Genotypic response varied widely across the environments as indicated by vectors that radiate in all the directions in PLS biplot that depicted genotypic variables, environments, and genotypes (Fig. 3 ). Variation in pattern of genotypic response within location is highly evident from the fact that V2T is independent of V1T, which was placed opposite to V1L. The biplot clearly differentiates normal and late-sown environments at most of the locations except at two contrasting environments: Jaipur in the first year and Varanasi in the second year. The response pattern of genotypes at Jaipur, relative to Varanasi, was featured by short crop duration, longer grain growth duration that supported better grain growth as indicated by relative position, and angles between the vectors in the PLS biplot.


Figure 3
View larger version (19K):
[in this window]
[in a new window]

  		Figure 3. Biplot based on partial least squares analysis of GxE interaction for 25 genotypes tested in 18 environments showing the relationship between genotype, environment, and genotypic variables. Genotypic variables are days to anthesis (DA), days to maturity (DM), plant height (PH), biomass at harvest (BM), thousand grain weight (TGW), grains per spike (GPS), grain weight per spike (GWS), grain growth duration (GGD), single grain weight (SGW), and harvest index (HI).

 
PLS analysis of genotypic variable and GxE interaction for grain yield placed genotypic variables such as days to anthesis, days to maturity and biomass in the right uppermost quadrangle, which was opposite to genotypic variables such as single grain weight, thousand grain weight, and grain growth duration (Fig. 3). Plant height and productive tillers were placed opposite to each other, indicating that many of the tall plant genotypes had less productive tillers. The variations in genotypic parameters are significant both in normal and warm environments. Genotypes such as 25 (WR 704), 11 (K 9162), 12 (Lok 1), and 14 (PBN 142) were placed in the quadrangle opposite to days to anthesis or days to maturity, indicating that they had early anthesis and/or maturity. In contrast to this, genotypes such as 2 (CBW 09), 3 (CBW 12), 15 (PBN 51), and 23 (WH 730) were in the same quadrangle as days to anthesis or days to maturity, indicating that they were featured by late flowering and/or maturity, more biomass, and short grain growth duration. Interestingly, genotypes 21 (Raj 4024), 16(PBW 497), and 10 (HW 2044), which displayed stable yield under late-sown conditions, were placed together, indicating that they share some common features that are responsible, such as a larger number of productive tillers and hence more biomass. Stable phenology was another feature indicated by the PLS biplot, as these genotypes were placed almost perpendicular to vectors of days to anthesis and grain growth duration.


   DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
This study attempted to address two specific questions, first, whether a genotype developed at a hot location could be suitable for other hot locations, and second, what are the most stable high-yielding genotypes that could be used as genetic stock for further improvement in grain yield under late-sown environments of the Indo-Gangetic plains of India, which are often featured by supraoptimal temperatures.

For these purposes, environments were classified by Ward's incremental sum of squares method into different groups based on temperature variables. Mean temperature for the preanthesis and postanthesis periods were considered so as to assess their relative significance to GxE interaction. Mean maximum temperatures of seventh week after planting were considered, as this was positively and significantly correlated with final grain yield of genotypes as revealed by preliminary correlation studies (not shown). Analysis of variance for temperature parameters revealed significant differences between the environment groups but not within the groups, indicating that they were distinct with respect to temperature regimes. It was clear from the analysis that all six environments that prevailed at Jaipur and Varanasi together were considerably hotter than environments at other locations. This can be largely attributed to proximity of Jaipur to the Thar Desert and delayed sowings at Varanasi. On the other hand, proximity to the Himalayas and sowing time (i.e., in early November) were the major reasons for cool weather in Environment Group I, represented by normal-sown conditions in K1T, L2T, K2T, and H1T. Differences in minimum temperatures at early stages of crop growth is a partial explanation for the differential genotypic response patterns at Jaipur and at Varanasi, as indicated by the PLS biplot for environmental variables and environments. During the second year, environments at Varanasi were influenced by mean minimum temperatures, in contrast to the environments in the first year, where influence of mean maximum temperatures was significant. This could be the major reason for contrasting patterns observed for crop phenology and grain weight in these two locations. It was also evident that the performance of genotypes was influenced by temperatures before as well as after the anthesis. Hence, superiority of a genotype under a hot environment was due not only to tolerance at the postanthesis stage but also at early stages of growth. These environments are comparable with the hot environments that exist in Australia and some part of USA (Moffat et al., 1990; Paulsen, 1994; Stone and Nicolas, 1994). In all the three groups of environments reported here, postanthesis mean temperatures were greater than 15°C, which is considered as the optimum for grain growth (Wardlaw and Wrigley, 1994). However, the Environment Group III was featured by more severe temperature stress with mean maximum temperatures greater than 35°C, which severely affects physiological processes at the time of grain-filling (Fischer, 1984; Stone and Nicolas, 1994; Stone et al., 1995; Ferris et al., 1998). There is good evidence that as few as 3 d of maximum temperatures above 35°C during grain-filling can cause considerable reduction in grain yield in wheat (Randall and Moss, 1990; Hawker and Jenner, 1993; Stone and Nicolas, 1994). In the present study, adverse effect of high temperatures on grain growth was highly conspicuous at Varanasi in the second year, during which crops were exposed to high temperature at both early and later stages of crop growth (Fig. 3 and 4 ).


Figure 4
View larger version (24K):
[in this window]
[in a new window]

  		Figure 4. Biplot based on partial least squares analysis of GxE interaction for 25 genotypes tested in 18 environments showing the relationship between genotype, environment, and environmental variables. Environmental variables are mean maximum temperatures of seventh week after planting (7WM), mean minimum temperature before anthesis (BAMN), mean maximum temperature before anthesis (BAMX), mean minimum temperature after anthesis (AAMN), and mean maximum temperature after anthesis (AAMX).

 
Grain yields at 18 different environments were analyzed for effects due to genotype and GxE interaction. This approach was used to understand the impact of environments on plant responses (mainly yield), to evolve suitable plant breeding strategies (Reynolds et al., 1998; Vargas et al., 1998; Lillemo et al., 2005). In the present study, SRG was used mainly because this regression model is suitable for grouping sites and cultivars without changing the ranks of cultivars. Further, the first two components of SRG are useful for summarizing and approximating a pattern of response that exists in the original data (Gabriel, 1971).

The SRG biplots revealed that contribution to GxE interaction for variability in grain yield was largely due to differential responses of genotypes to environments at K1T, K2L, L1L, and V1T, which represented a relatively cool environment, though this included timely and late-sown environments. This can be attributed to the fact that the set of genotypes chosen for this study included those targeted for both timely and late-sown environments, and these genotypes were expected to have different levels of adaptability to the environments. Discrimination of genotypes on the basis of grain yield appeared to be less in hot locations such as Varanasi and Jaipur, relative to other locations, which were relatively cool, indicating that thermal stress must have reduced the variation in responses of genotypes studied in this experiment. The biplot clearly revealed that response patterns of genotypes for timely and late-sown conditions were nearly similar at many of the locations, except at Varanasi and Jaipur, where even the normal-sown environment was featured by high temperatures. Thus SRG analysis was able to differentiate hot environments from cooler ones on the basis of GxE interaction for grain yield across the environments. It was also evident that though Varanasi and Jaipur were hotter than other locations, these two locations had different patterns of genotypic response. This was also confirmed by PLS biplot (Fig. 4), which could partially explain this on the basis of variation in impact of mean minimum temperatures between as well as within the locations. This suggests that probably the possibility of a common strategy to evolve thermal-stress-tolerant genotypes for these two hot locations is feasible only if stability of genotypes is prioritized in the breeding program. Although Jaipur was characterized by high temperature, two of its genotypes, 20 (Raj 40104) and 21(Raj 4024), displayed low stability. However, two other lines, 17 (Raj 3765) and 22 (Raj 4027), of Jaipur displayed good performance for grain yield in the stability analysis across 18 environments. Further, genotypes developed by this center (except Raj 4027) did not perform so well in the late-sown conditions of the other hot location, Varanasi, which had contrasting temperature regimes. Relative humidity has been cited as one of the important parameters that influence crop growth in a hot environment (Reynolds et al., 1998; Vargas et al., 1998). This may partially explain the differential pattern of genotypes observed at Jaipur, which is featured by hot and dry weather, in contrast to Varanasi, which is hot and humid. Further, nights are hotter at Varanasi than at Jaipur, as revealed by mean minimum temperature vectors in PLS biplot. These observations support our suggestion that germplasm exchange between these locations should be based on stability of genotypes as indicated by biplots. Selection based on yield potential appears to be feasible for relatively less hot environments such as Hisar and Karnal. The relatively low yield of some of the more stable genotypes across locations suggests that they need to be crossed with high-yielding genotypes to combine heat tolerance and yield. Limited backcrossing combined with a shuttle breeding approach, as practiced at CIMMYT (Singh et al., 1998, 2000; Singh and Huerta-Espino, 2004), would be highly desirable to combine heat tolerance with higher yield in the background of these genotypes. The combined analysis of data from 18 environments suggests six genotypes, 2 (CBW 09), 3(CBW 12), 9 (HUW 543), 17 (Raj 3765), 18 (Raj 4000), and 22 (Raj 4027), as more stable than others for grain yield. Under late-sown conditions, three genotypes, 16 (PBW 497), 17(Raj 3765), and 10 (HW 2004), displayed the highest yield with a low to moderate decline in comparison to timely sowing (Table 6). However, the lowest decline due to late sowing across environments was displayed by two other genotypes, 25 (WR 704) and 5 (HD 2285), which had relatively low grain yield. This suggests that 25 (WR 704) and 5 (HD 2285), appear to possess relatively higher thermal stress tolerance than other genotypes tested. Heat tolerance in these genotypes could be due to early flowering, long grain growth duration, and tendency to escape from high temperatures, as indicated by PLS biplot for genotypic variables (Fig. 3). Both of these genotypes possessed almost similar days to anthesis and maturity in comparison to 7 (HUW 234) and 12 (LOK 1). Genotype 7 (HUW 234) was released for late-sown conditions of northeastern Plains zone of India in 1986 and is still the most dominant cultivar in the eastern Gangetic plains, especially under late-sown conditions, due to its excellent stability across varying soil and management environments and capacity to tolerate terminal heat stress. The other cultivar, 12 (LOK 1), was released for cultivation in 1982 for the central zone of India, which is also considered to be an area having warmer temperatures coupled with low availability of moisture. These two genotypes are being used widely by different breeding programs including that of CIMMYT, Mexico, for breeding short-duration and terminal-heat-stress-tolerant genotypes for South Asia. The present study establishes that there are genotypes possessing superior stability to these two popular cultivars not only under late-sown conditions but also under timely sown conditions. Therefore, the superior genotypes 17 (Raj 3765) and 22 (Raj 4027) identified in this study should be used in breeding programs targeting tolerance to high-temperature stress. Other genotypes, specifically 21 (Raj 4024), 16 (PBW 497), and 10 (HW 2044), that displayed superior and stable yield under late-sown condition across different environments should also be used as donor parents in the ongoing breeding programs to develop high-temperature-tolerant wheat varieties that can survive the growing threat of global warming. In earlier studies 3 (CBW 12) was found to have higher grain growth rate per unit growing degree day than any other genotype mainly because of enhanced maturity and short duration under late-sown conditions (Rane and Chauhan, 2002a). The present study reveals that such genotypes could be useful for moderately high temperature environments or favorable environments, as their stable and superior performance appear to be governed by yield potential in nonstressed environments. Grain weight per spike recorded in 13 (NIAW 845) and 25 (WR 704) was more than that of other genotypes, which were raised in polythene tunnels where temperatures were very high (Rane and Chauhan, 2002b). These genotypes were also found to have ability to mobilize more stem reserves for grain growth (Rane et al., 2003). However, grain yields of these genotypes were low and highly unstable across the environments mainly because of few effective tillers and crop lodging. It is suggested that with appropriate breeding strategies these stress-tolerant genotypes can be successfully used to introgress thermal tolerance into high-yielding stable genotypes for developing thermotolerant wheat suitable for hot environments of the Indo-Gangetic Plains.

This study also reveals that only some of the genotypes developed at a hot location such as Jaipur were useful for another hot environment, at Varanasi, where contrasting temperature regimes differentiated genotypes in a different pattern. Hence, a common research strategy with emphasis on grain yield stability may be highly feasible for developing heat-tolerant genotypes. Genotypes Raj 3765 and Raj 4027 with stable grain yield across different locations may serve as superior genetic stocks to develop genotypes tolerant to high temperatures that prevail commonly under late-sown conditions. Genotypes with high grain yield potential can perform better under moderately hot environments that prevail at relatively cooler locations such as Karnal, Hisar, and Ludhiana, which are located in the northwestern part of the Indo-Gangetic plains.


   ACKNOWLEDGMENTS
 
Financial support from the Indian Council of Agricultural Research through an NATP-CGP grant and assistance from technical staff in respective locations are gratefully acknowledged. We also thank three anonymous reviewers for critically reviewing this manuscript and for their constructive comments.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permission for printing and for reprinting the material contained herein has been obtained by the publisher.

Received for publication July 20, 2006.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Rane, J.
Right arrow Articles by Joshi, A. K.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Rane, J.
Right arrow Articles by Joshi, A. K.
Agricola
Right arrow Articles by Rane, J.
Right arrow Articles by Joshi, A. K.
Related Collections
Right arrow Field evaluation techniques
Right arrow Plant and Environment Interactions
Right arrow Crop Ecology

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Vadose Zone Journal
Journal of Natural Resources
and Life Sciences Education		 Soil Science Society of America Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome