Breeding for increased water use efficiency in chickpea

BREEDING FOR INCREASED WATER USE EFFICIENCY IN CHICKPEA By Peter Kaloki MSc. Plant Breeding and Genetics A thesis submitted for the fulfilment of the requirements for Doctor of Philosophy Faculty of Agriculture and Environment Plant Breeding Institute, Cobbitty The University of Sydney March, 2017 i Certificate of Originality I certify to the best of my knowledge that all the content of this thesis is my own work and that it has not been submitted elsewhere for the purposes of obtaining a diploma or degree in any other University. I also certify that the thesis content is the product of my own work and that all sources that assisted in preparing this thesis have duly been acknowledged. Peter Kaloki ii ABSTRACT Traditionally, chickpea is grown as a winter crop with in-crop rain or stored soil moisture, or as a spring crop using residual stored soil moisture. In the semi-arid tropics, it is grown when rainfall is tapering off in the late rainy season and utilises moisture stored in the soil profile. These growing conditions are characterised by a gradual decline in soil moisture towards the end of the growing season leading to terminal drought. Drought causes up to 50% yield losses in chickpea, however, depending on the genotype, environment, and type of drought experienced, seed yield losses can range from 30 – 100%. The effect of drought will be exacerbated by global warming which is projected to be responsible for a 20% increase in water shortages in drought prone areas. Since 80% of the world’s allocable water is consumed in irrigated agriculture, and water resources for agriculture are generally decreasing, it may not be feasible to grow chickpeas under irrigation to mitigate the effect of drought. Breeding cultivars with high water use efficiency (WUE) is a more practical and economical long-term approach to increasing yields in drought prone areas. WUE leads to moderate water uptake while maintaining increased yields under drought conditions making WUE an integral part of breeding programs. Any modifications above the soil surface have an effect on WUE since it impacts on the soil water balance via soil water evaporation and infiltration. This necessitates the incorporation of management practices, such as tillage, in studies analysing WUE. Since WUE is a complex trait, secondary traits that are easy to measure and that have genetic variation, high heritability and are associated with yield under water-limited conditions make breeding for WUE easier. Little attention has been paid to the pattern of water use in legumes and the relationship between water used, WUE and seed yield. Despite evaluation of WUE in chickpea in various studies, little has been achieved as those studies focused on single factors affecting WUE, which caused variability in outcomes due to a failure to integrate other factors. The central research question of this study was: can chickpea yields be sustained by increased water use efficiency under drought conditions? The aims of this thesis were to study the genetic variation underpinning WUE and grain yield in different tillage and irrigation regimes, as well as the basis of yield formation under water limited conditions. Water use and WUE are important traits under water-limited conditions. It was hypothesised that genotypes with high WUE would produce high yields under water-limited conditions. For this hypothesis to be tested, a total of 36 entries were planted in the field at the IA Watson Plant iii Breeding Institute, The University of Sydney in Narrabri, northwest New South Wales in Australia. Water use was monitored using a neutron probe moisture meter and WUE calculated using the soil water balance method. Grain yield was higher under irrigation (1722 kg ha-1) than rainfed conditions (1478 kg ha-1). No till plots resulted in an average yield of 1658 kg ha1 which was 7.4% higher than in the till regime. There were no significant differences in water use; however, there were significant differences for WUE. WUE was higher under no till (5.02 kg ha-1 mm-1 than under till (4.87kg ha-1 mm-1), and higher under irrigation (5.05 kg ha-1 mm-1) than under rainfed conditions (4.84 kg ha-1 mm-1). Sonali was the highest yielding genotype and also had the best WUE. Identifying drought tolerant genotypes to be used as sources of tolerance in a breeding program is imperative. Traits that can confer drought tolerance under field conditions should be considered instead of yield alone. It was hypothesised that drought selection indices differ in their prediction accuracy and that some indices can be used to predict marker traits that can confer tolerance to drought in the field. To test this hypothesis, phenological, morphological, physiological and yield component data were analysed from the experiments performed in Narrabri. Drought indices were calculated and multiple linear regression was used to identify the most important traits that explained variation in yield. The stress tolerance index, mean relative performance and relative efficiency index were highly and positively associated with yield. These three traits were identified as the most effective indices for use in chickpea using principal component analysis compared with drought resistance index, yield index and yield stability index, which were not as suitable. Sonali, ICCV 96853 and PBA Slasher were identified as drought tolerant genotypes whereas Amethyst and Genesis 079 were identified as susceptible to drought. A total of 21 traits (Agyeman et al., 2015) out of 40 were identified as important in drought tolerance. The indices identified normalised difference vegetation index (NDVI) at early podding and late podding, as well as chlorophyll content at late podding, as useful marker traits to identify genotypes with potentially high yield and high drought tolerance. Sustaining yield under different environments is important for the grower as well as the plant breeder. Genotype by environment interaction affects varietal ability to sustain yields across environments. It was hypothesised that there would be a significant genotype by environment interaction and hence, yield would not be stable across environments. To test this, 36 genotypes were sown using a two factorial experimental design in two seasons under no till, with and without irrigation, and till, with and without irrigation, in Narrabri making a total of eight iv environments. The data were analysed using restricted maximum likelihood (REML) to check for genotype by environment interaction as well as genotype and genotype by environment interaction (Staggenborg and Vanderlip) biplot analysis to identify stable and high yielding genotypes. There was a significant genotype by environment interaction and genotype performance varied with environment. Generally, the yields in 2014 were higher than those in 2015 with 58% of the variation in yield accounted for by the year (season) effect. No till with irrigation in 2014 resulted in the highest average yield and till rainfed in 2015 resulted in the lowest mean yield. Some genotypes were more stable and high yielding than others. PBA Slasher and ICCV 96853 were high yielding and stable, whereas Genesis 079 was high yielding and very unstable. Sonali and Amethyst had moderate stability. The plant ideotype approach is an alternative strategy to empirical breeding and allows the breeder to predict the ideal genotype in the target environment. Each ideotype is designed to grow in a defined target environment, hence, it is important to characterise the environment. It was hypothesised that selecting for key plant traits can confer drought tolerance and that abiotic stress sensitivity varies across plant phenophases. To test these hypotheses, data generated from the Narrabri field experiment was used. The key phenological, morphological and physiological traits were determined for ideotype targeting using multiple linear regression and ideotype values assigned depending on trait relationship with yield and other traits. The ideotype was then tested against selected commercial varieties (Sonali, PBA Hattrick, Kyabra, Tyson and Amethyst) in silico in the Australian grain belt using the APSIM crop model. The constructed chickpea ideotype showed 76% resemblance to Sonali which performed well under water-limited conditions. Simulated yield ranged from 760 to 3902 kg ha-1 across the Australian grain belt, with consistently higher yield in the ideotype compared with the commercial cultivars. The growing environments were grouped into three major clusters using the soil water deficit method with varying water stress levels. Grain filling is the most critical stage where soil moisture deficit caused chickpea yield loss. By incorporating key target traits and targeting the right environment, chickpea yields can be sustained. This study shows that there is genetic variation for WUE and it is a major component of drought tolerance. By identifying drought tolerant genotypes which are high yielding and stable, yields may be sustained under water limited conditions. By targeting a chickpea ideotype for specific environments, plant breeders can have a more focused strategy and hence, faster delivery of technologies to develop cultivars that are suitable for the target environment v Table of contents ABSTRACT ............................................................................................................................. iii Table of contents ....................................................................................................................... vi List of Figures ............................................................................................................................ x List of Tables ......................................................................................................................... xiii List of Abbreviations ............................................................................................................... xv Acknowledgements ................................................................................................................. xvi 1.0 GENERAL INTRODUCTION ....................................................................................... 1 2.0 LITERATURE REVIEW .................................................................................................... 5 2.1 Introduction ...................................................................................................................... 5 2.2 Origin and cytogenetics .................................................................................................... 5 2.3 Distribution, climate, area, production and uses of chickpea........................................... 6 2.4 Mode of reproduction and types of chickpea ................................................................... 8 2.5 Chickpea genetic resources .............................................................................................. 8 2.6 Chickpea production constraints ...................................................................................... 9 2.7 Chickpea cropping systems and tillage practices ........................................................... 10 2.8 Conservation agriculture and its implication on chickpea cultivation ........................... 11 2.9 Drought resistance mechanisms in plants ...................................................................... 13 2.10 Effects of water deficit on chickpea growth and development .................................... 14 2.11 Chickpea physiological responses to water deficit ...................................................... 15 2.12 Water use efficiency and associated breeding efforts .................................................. 17 2.13 Breeding for increased water use efficiency using surrogates ..................................... 19 2.14 Phenotyping target physiological traits in chickpea ..................................................... 20 2.14.1 Canopy temperature ............................................................................................... 21 2.14.2 Plant vigour and plant green biomass .................................................................... 22 2.14.3 Photosynthetically active radiation (PAR) ............................................................ 23 2.14.4 Chlorophyll content ............................................................................................... 23 2.15.5 Root traits .............................................................................................................. 24 2.15 Chickpea ideotype development .................................................................................. 24 2.16 Crop modelling and ideotype design ............................................................................ 26 2.17 Conclusion.................................................................................................................... 27 CHAPTER 3: GENERAL MATERIALS AND METHODS .................................................. 28 3.1 Introduction .................................................................................................................... 28 3.2 Experimental site ............................................................................................................ 28 vi 3.3 Experimental design ....................................................................................................... 28 3.4 Field experiment sowing ................................................................................................ 29 3.5 Field agronomic practices .............................................................................................. 30 3.6 Data parameters .............................................................................................................. 30 3.7 Field irrigation ................................................................................................................ 30 3.8 Weather data ................................................................................................................... 30 3.9 Data analysis .................................................................................................................. 31 CHAPTER 4: WATER USE, WATER USE EFFICIENCY AND YIELD VARIATION IN CHICKPEA GENOTYPES ..................................................................................................... 32 4.1 Introduction .................................................................................................................... 32 4.2 Materials and methods ................................................................................................... 33 4.3 Results ............................................................................................................................ 36 4.3.1 Precipitation and temperature .................................................................................. 36 4.3.2 Seed yield ................................................................................................................ 37 4.3.3 Seed yield variation and interaction under different tillage and moisture regimes . 40 4.3.4 Water use ................................................................................................................. 41 4.4.5 Water use efficiency under different tillage and moisture regimes (individual analysis) ............................................................................................................................ 42 4.3.6 Water use efficiency under different tillage and moisture regimes (combined analysis) ............................................................................................................................ 44 4.3.7 Genetic variation for water use and WUE under different tillage and moisture regimes.............................................................................................................................. 47 4.3.8 Water use, water use efficiencyand yield relationships under rainfed and irrigated conditions.......................................................................................................................... 48 4.3.9 Heritability and genetic advance of WUE ............................................................... 49 4.3.10 Genotypic, phenotypic and environment coefficient of variation for WUE under different tillage and moisture regimes .............................................................................. 50 4.4 Discussion ...................................................................................................................... 50 4.5 Conclusions .................................................................................................................... 52 CHAPTER 5: THE BASIS OF CHICKPEA YIELD FORMATION UNDER WATER LIMITED FIELD CONDITIONS. .......................................................................................... 53 5.1 Introduction .................................................................................................................... 53 5.2 Materials and methods ................................................................................................... 55 5.3 Results ............................................................................................................................ 57 vii 5.3.1 Phenological, morphological and physiological traits for yield formation under water stressed conditions .................................................................................................. 57 5.3.2 Grain yield and drought indices............................................................................... 59 5.3.3 Grain yield relationships under well-watered and water stressed conditions .......... 61 5.3.4 Correlation analysis for grain yield and drought indices ......................................... 61 5.3.5 Selection of the best drought tolerance index for chickpea ..................................... 63 5.3.6 Effect of water deficit on important traits associated with chickpea grain yield under water limited conditions ......................................................................................... 64 5.3.7 Associations between trait relationships and chickpea drought indices .................. 66 5.4 Discussion ...................................................................................................................... 67 5.5 Conclusions .................................................................................................................... 71 CHAPTER 6: EFFECT OF GENOTYPE BY ENVIRONMENT BY MANAGEMENT INTERACTIONS ON CHICKPEA PHENOTYPIC STABILITY ......................................... 72 6.1 Introduction .................................................................................................................... 72 6.2 Materials and methods ................................................................................................... 74 6.3 Results ............................................................................................................................ 75 6.3.1 Weather data ............................................................................................................ 75 6.3.2 Grain yield under different environments ............................................................... 76 6.3.3 Factors accounting for grain yield variation ............................................................ 77 6.3.4 Genotype, environment and genotype by environment interaction ......................... 78 6.3.5 Test environment evaluation ................................................................................... 79 6.3.6 The ideal test environment ...................................................................................... 80 6.3.7 Mean grain yield performance and stability test ..................................................... 82 6.3.8 Selecting the ideal genotype .................................................................................... 83 6.3.9 Mega-environment analysis ..................................................................................... 84 6.4 Discussion ...................................................................................................................... 85 6.5 Conclusions .................................................................................................................... 87 CHAPTER 7: DEVELOPMENT OF A DROUGHT TOLERANT CHICKPEA IDEOTYPE FOR THE AUSTRALIAN GRAIN BELT .............................................................................. 88 7.1 Introduction .................................................................................................................... 88 7.2 Materials and methods ................................................................................................... 90 7.2.1 Field experiments .................................................................................................... 90 7.2.2 Chickpea ideotype development .............................................................................. 90 7.2.3 Environmental characterisation: the soil water deficit approach ............................. 91 viii 7.3 Results ............................................................................................................................ 92 7.3.1 Chickpea ideotype ................................................................................................... 92 7.2 Validation of the APSIM-Chickpea model .................................................................... 96 7.3 Simulated yield ............................................................................................................... 97 7.4 Environmental characterisation and soil water deficit patterns ..................................... 97 7.5 Stress timing and the critical period for yield penalty ................................................. 101 7.6 Discussion .................................................................................................................... 103 7.7 Conclusions .................................................................................................................. 104 8.0: GENERAL DISCUSSION ............................................................................................. 106 8.1 Introduction .................................................................................................................. 106 8.2 Water use and water use efficiency in chickpea .......................................................... 110 8.3 Chickpea yield under water limited conditions ............................................................ 110 8.4 Chickpea phenotypic stability ...................................................................................... 111 8.5 Chickpea ideotype ........................................................................................................ 111 8.6 Summary of discussions ............................................................................................... 112 8.7 Conclusions .................................................................................................................. 112 8.7 Further research ............................................................................................................ 113 References .............................................................................................................................. 114 ix List of Figures Chapter 2 Figure 2.1: Climate data from 1981 to 2010 for chickpea growing areas in Australia….. 6 Figure 2.2: Some target traits for chickpea physiological breeding……………………… 21 Figure 2.3: Schematic illustration of physiological breeding. ………………………….. 26 Chapter 3 Figure 3.1: Narrabri experimental field layout in 2014 and 2015 for chickpea water use 29 efficiency experiments…………………………………………………………………... Chapter 4 Figure 4.1: Neutron probe moisture meter ……………………………………………… 34 Figure 4.2: Neutron probe access tube…………………………………………………... 34 Figure 4:3: Daily maximum and minimum temperatures for 2014 (a) and 2015 (b) and rainfall for 2014 (c) and 2015 (d) at the water use efficiency experimental field site: Narrabri………………………………………………………………………………….. 37 Figure 4.4: Mean chickpea grain yields under irrigation and rainfed conditions (a), no till and till systems (b), and different seasons (c)……………………………………… 38 Figure 4.5: Mean chickpea water use under no till and till systems in 2014 (a) and 2015 (c), and under irrigation and rainfed conditions in 2014 (b) and 2015 seasons (d)…….. 41 Figure 4.6: Mean volumetric water content (denoting root water access) at various soil depths during the growing season in 2014 (a) and 2015 (b)…………………………….. 42 Figure 4.7: Chickpea WUE under different tillage (1), moisture (2), season (3) and tillage by moisture interaction (4)...…………………………………………………. 47 Figure 4.8: Relationships between water use, water use efficiency and yield ………….. 50 Chapter 5 Figure 5.1: Relationship between irrigated (well-watered) and rainfed (water stress) yield for the different chickpea genotypes analysed for drought tolerance………………….. 61 Figure 5.2: Linear regression of chickpea grain yield against six drought indices……… 62 x Figure 5.3: Principal component scatter plot for chickpea genotypes and drought indices……………………………………………………………………………………. 64 Chapter 6 Figure 6.1: Chickpea grain yields for different tillage and moisture regime 77 environments over two years (2014 and 2015)………………………………………….. Figure 6.2: Environment scatter plot for evaluation of the test environment and chickpea genotypes……………………………………………………………………… 80 Figure 6.3: Test environment comparison scatter plot for evaluating genotype and 81 environment interactions in chickpea yield……………………………………………... Figure 6.4: Principal component analysis scatter plot for evaluating grain yield performance and stability in chickpea genotypes……………………………………….. 83 Figure 6.5: Scatter plot for evaluating the ideal chickpea genotype…………………….. 84 Figure 6.6: Mega environment scatter plot for evaluating chickpea yield across environments…………………………………………………………………………… 85 Chapter 7 Figure 7.1: Flow diagram for chickpea ideotype construction………………………… 91 Figure 7.2: Evaluation of chickpea for drought tolerance using minimum spanning tree 96 for genotype similarity…………………………………………………………………. Figure 7.3: Evaluation of chickpea traits using APSIM modelling. (a) Days to 50% flowering (days) and (b) grain yield (kg ha-1) for observed and simulated data……… 96 Figure 7.4: Evaluation of chickpea yields across different production environments ….. 97 Figure 7.5: Dendrogram of Australian chickpea production environment characterisation based on soil moisture deficit. Arrows indicate the start of a new cluster or group…………………………………………………………………………. 98 Figure 7.6: Evaluation of APSIM-predicted chickpea traits for drought tolerance based on soil water deficit clusters. (a) Mean days to 50% flowering (grey bars) and maturity (black bars) and (b) mean grain yield (kg ha-1). ………………………………………... 100 Figure 7.7: Frequency predictions (%) for chickpea yield based on cluster groupings 101 xi Figure 7.8: APSIM-predicted soil water deficits in different growth stages over a 100 year period x 50 locations x six varieties………………………………………………... 102 Chapter 8 Figure 8.1: Schematic presentation of the scope, aims and findings of the present study 107 xii List of Tables Chapter 2 Table 2.1: Water use efficiency of chickpea genotypes across different environments…... 19 Chapter 3 Table 3.1: List of chickpea genotypes for water use efficiency experiments at Narrabri…. 29 Chapter 4 Table 4.1: Mean chickpea seed yield (kg ha-1) under different management and experimental conditions………………………………………………………………….. 38 Table 4.2: Wald statistic for main effects (tillage, moisture, genotype, season) and their interaction on chickpea seed yield………………………………………………………… 40 Table 4.3: Mean chickpea WUE among genotypes under different tillage, irrigation and seasonal conditions………………………………………………………………………… 43 Table 4.4: Mean chickpea WUE for combined analysis in 2014 and 2015……………….. 45 Table 4.5: Variation for chickpea water use among genotypes under different tillage and moisture regimes…………………………………………………………………………... 47 Table 4.6: Components of variation in chickpea WUE in 2014 and 2015 48 Table 4.7: Heritability estimates and genetic advance for chickpea genotypes under different tillage and moisture regimes……………………………………………………. 50 Table 4.8: Coefficient of variation for WUE in different tillage and moisture regimes 50 Chapter 5 Table 5.1: Drought tolerance indices for evaluating chickpea yield under water-limiting conditions …………………………………………………………………………………. 57 Table 5.2: Traits explaining variation in chickpea yield under water stressed conditions, 58 correlations with grain yield, heritability and genetic advance…………………………… Table 5.3: Grain yield and drought tolerance indices for chickpea genotypes grown under well-watered and water stressed conditions……………………………………………….. 60 Table 5.4: Trait means and per cent change due to water deficit in chickpea genotypes…. 65 Table 5.5: Correlation between drought indices and important traits in chickpea grown 67 under water deficit conditions……………………………………………………………… xiii Chapter 6 Table 6.1: Field environments with different tillage and moisture regimes for analysis of chickpea phenotypic stability 74 Table 6.2: Average weather conditions for each environment experienced by chickpea 75 genotypes analysed for phenotypic stability………………………………………………. Table 6.3: The main factors accounting for grain yield variation in chickpea grown across different environments……………………………………………………………... 78 Table 6.4: Combined analysis of variance (ANOVA) for genotype and environment 79 effects on mean chickpea grain yields…………………………………………………….. Chapter 7 Table 7.1: Wald statistic, correlations and decisions used to construct the chickpea ideotype …………………………………………………………………………………… 93 Table 7.2: Trait range, genotype and ideotype values for evaluating chickpea drought tolerance through APSIM modelling……………………………………………………… 94 Table 7.3: Australian chickpea production environmental clusters based on soil water deficit………………………………………………………………………………………. 98 Table 7.4: Multiple linear regression of various growth stages in relation to chickpea yield……………………………………………………………………………………….. 102 Chapter 8 Table 8.1: Thesis summary with objectives, key findings and outcomes………………… 108 xiv List of Abbreviations ATP Adenosine triphosphate APSIM Agricultural Production Systems Simulator CGIAR Consultative Group on International Agricultural Research CID Carbon isotope discrimination CO2 Carbon dioxide DI Drought resistance index ECV Environmental coefficient of variation GA Genetic advance GCV Genotypic coefficient of variation GEI Genotype by environment interaction GGE Genotype plus genotype by environment interaction GLM Generalised linear models ICARDA International Centre for Research in to Dry Areas ICRISAT International Crops Research Institute for the Semi-Arid Tropics LAI Leaf area index MET Multi-environment trials MRP Mean relative performance NDVI Normalised difference vegetation index NIR Near infrared NVT National variety trials PAR Photosynthetic active radiation PC Principal component PCV Phenotypic coefficient of variation REI Relative efficiency index REML Restricted maximum likelihood STI Stress tolerance index USDA United States Department of Agriculture VPD Vapour pressure deficit WS Water stress WUE Water use efficiency WW Well watered YI Yield Index xv Acknowledgements I would like to thank almighty God, because this far, He has brought me. I would like to express my sincere gratitude to my supervisor, Associate Professor Daniel Tan, University of Sydney for his valuable advice and input during my studies. His speedy response and encouragement kept me going. Special thanks to my supervisor Professor Richard Trethowan, Director IA Research Institute, University of Sydney, Narrabri for his support and guidance throughout my studies. Whenever I thought I am stuck, he always had a solution. Special thanks also to my auxiliary supervisor Dr Helen Bramley, University of Sydney for her great advice and in-depth academic discussions. Special thanks to Dr Qunying Luo for her guidance and advice on APSIM modelling. I would like also to express my gratitude to the entire staff of the Plant Breeding Institute, both in Narrabri and Cobbitty for their support and help over the years. Special thanks to Angela Pattison, Phil Davies, Annette Tredea, Peter Bell, Kate Rudd, Melissa Eather, Anthony Vuragu and James Bell for going an extra mile to help me during the course of my studies. I am also very grateful to Dr Chong Mei for helping with DNA extraction and Dr Urmil Bansal for the many discussions and advice on molecular biology. I am also grateful to Dr Pooran Gaur, Principal scientist in chickpea breeding ICRISAT for his support, encouragement and advice over the years. Special thanks to Dr Said Silim, former Director ICRISAT eastern and southern Africa for his support and shaping my career in plant breeding. I am grateful to all my fellow students for the academic discussions we had that gave us deeper insights into our field of study and also for the encouragement and support over the years. I would like to express my sincere gratitude to my wife Elizabeth Kiilu for supporting and encouraging me throughout the course of my studies. Special thanks to our daughters Sydney, April and Summer for always putting a smile on my face even when things seemed to be tough. Last but not the least, I would like to acknowledge the Australian government for sponsoring my education through the Australia Awards Africa scholarship. Special thanks to the scholarship administrators at the University of Sydney (Amy, Sue, Bojan and Annie) for their support over the years. xvi SUBMITTED PUBLICATION FROM THIS THESIS Chapter 7 Kaloki, P., Luo, Q., Trethowan, R. Tan, D. Can the development of drought tolerant ideotype sustain Australian chickpea yield? Agricultural and Forest Meteorology Journal. Date submitted; 10 January 2017. Reviewed and re-submitted on 11 August 2017. CONFERENCE PRESENTATION FROM THIS THESIS Chapter 5 Kaloki, P., Trethowan, R. Tan, D. Target traits for improving water use efficiency in chickpea under water limited environments. University of Sydney, Faculty of Agriculture Symposium, 12 July 2016, Sydney, Australia. POSTER PRESENTATIONS FROM THIS THESIS Chapter 7 Kaloki, P., Trethowan, R. Tan, D. Chickpea plant ideotype development for semi-arid subtropical climates to assist plant breeding. Synergy in Science: Partnering for Solutions. ASA, CSSA, SSSA International Annual Meetings, 15-18 November 2015, Minneapolis, Minnesota, U.S.A. xvii 1.0 GENERAL INTRODUCTION Chickpea (Cicer arietinum L.), a grain legume, has been a focus crop in recent times with renewed interest in its cultivation due to its high protein content and soil amelioration capabilities. The average annual area under cultivation between 2010 – 2012 was 12.4 million ha (FAOSTAT, 2012), which was spread across 52 countries in the Indian sub-continent, Mediterranean basin, Australia, East Africa and the Americas. These areas lie under the tropics, subtropics with winter rainfall and subtropics with summer rainfall as described by Kassam (1981).Being a cool season crop, chickpea cultivation has been traditionally restricted to cool climates. This implies that they grow well under cool temperatures during their vegetative stage and as they change to reproductive phase, temperatures start to increase. However, during the flowering phase, temperatures lower than 14°C to 16°C cause flower abortion (Berger et al., 2004). As a result of breeding efforts coupled with its agronomic benefits, the growing area has since expanded to include the semi-arid tropics, where it has become a main food security (staple) crop in the drought prone areas. In these areas, they are cultivated such that the vegetative phase coincides with cool temperatures. Drought is the most limiting abiotic factor during various chickpea growth phases (Gunes et al., 2008, Boyer, 1982). It can be either intermittent and occasioned by a break in the normal rainfall pattern during the growing season, resulting in insufficient rainfall overall, or terminal drought resulting from continued moisture decline from the soil profile towards the end of the growing season (Canci and Toker, 2009). In all environments where chickpea is grown, terminal drought is almost certain (Turner, 2003) accounting for up to 50% of chickpea production losses (Varshney et al., 2013b). However, seed yield losses vary depending on the genotype, the type of drought experienced, and the environment, and can range from 30% to 100% (Leport et al., 1999). This situation is expected to be exacerbated by climate change, where it has been predicted that there will be more frequent drought events due to a general reduction in the amount of rainfall in the arid and semi-arid areas (IPCC, 2007). Chickpea morphology, phenology and physiology are affected by drought in various ways. The most sensitive growth stage to water deficit is at flowering and early podding (Khanna-Chopra and Sinha, 1987). Early transient water deficit has been shown to reduce flower production by almost 50%, increase flower abortion and reduce pod abortion compared with well-watered 1 controls in two chickpea cultivars, Rupali, a desi-type and Almaz, a kabuli type (Fang et al., 2011). Terminal drought caused 33-63% flower reduction, 37-56% flower abortion and 5473% pod abortion, in the same cultivars (Fang et al., 2010). Behboudian et al. (2001) reported that pod formation was greatly reduced after moisture stress was induced, although if induced at late flowering, it had minimal effect on pod production. The total number of pods was reduced by 66-75% in plants exposed to early podding moisture stress compared with the well-watered control (Leport et al., 2006). Pods formed before water stress was imposed were not affected in terms of dry mass, whereas those formed later had their final dry mass reduced (Behboudian et al., 2001). Although pod abortion was increased under increased moisture stress, seed abortion and individual seed mass were not (Behboudian et al., 2001). Davies et al. (1999) reported that chickpeas exposed to terminal drought under field conditions had a shorter seed filling duration and seed filling rate, resulting in smaller final seed size. Terminal drought reduced seed yield by 58-95% compared with the irrigated controls (Leport, 1999; Leport 2006 in Fang 2011). However, early transient water deficit in a pot study increased the rate of seed filling and final seed size at maturity compared with the well-watered control (Fang et al., 2011). Moisture stress at early podding reduced the number of seeds per pod from two to one in kabuli whereas desi-types were not affected (Leport et al., 2006). Fewer seeds per pod and smaller seed size caused a decrease in seed yield (Leport et al., 1999, Fang et al., 2010). Moisture stress induced at early podding caused a reduction in seed size and seed yield by 28% and 90%, respectively, although moisture stress at late podding did not reduce the seed size (Leport et al., 2006). Reduction in seed yield under the early transient water deficit was lower compared with terminal yield losses (Fang et al., 2011). Drought tolerance research has been very difficult (Tuberosa and Salvi, 2006) primarily due to a lack of proper understanding of the physiological basis of yield under drought conditions, as well as its quantitative inheritance nature (Sinclair, 2011). One of the key steps for a breakthrough in drought tolerance research is an understanding of the physiological basis of drought, which will in turn, open new frontiers in molecular breeding strategies (Reyazul et al., 2012). The multifaceted nature of drought needs a more comprehensive approach and deeper understanding of all its components. It is, therefore, prudent to dissect yield under drought conditions, which in effect is a function of water uptake, water use efficiency and harvest index 2 (Passioura, 1977). Little attention has been paid to the pattern of water use in legumes and the relationship between water used and seed yield (Zhang et al., 2000b). Despite evaluation of water use efficiency (WUE) in chickpea in various studies, little progress has been achieved because those studies focused on single factors affecting WUE. This causes variability in results from different studies due to a failure to integrate the various factors responsible for WUE (Gan et al., 2010). The identification of key morphological, physiological and biochemical traits that are associated with stress tolerance is important in understanding plant responses to water deficit conditions (Araus et al., 2002, Poormohammad et al., 2007, Condon et al., 2004, Reynolds et al., 1999). Once properly identified, these morphological, physiological and biochemical responses may be used as surrogates to select for WUE. This can be carried out in the framework of target trait based breeding, which has gained primacy in recent years as opposed to general breeding for increases in yield. Plant breeders are using easily measurable traits as surrogates for traits that were traditionally difficult to breed for. Drought tolerance is a very complex trait and WUE, which is one of the major components of drought adaptation, is complex as well. Hence, it is of prime importance to improve other traits that give an additive gene effect to eventually increase WUE and drought tolerance. Breeding cultivars with high WUE is a more practical and economical approach to improving yields in drought prone areas (Yong'an et al., 2010). The genotype and crop management practices play a key role in plant-water interactions. Hence, the need to understand more about genotype by environment by management interactions. More recently, chickpea has increasingly been cultivated under zero or minimum tillage systems, coupled with retention of crop residues on the soil surface, to conform to the principles of conservation agriculture (Bimbraw, 2016, Hobbs, 2007). Any modifications above or on the soil surface have an effect on water use efficiency since it impacts on the soil water balance via soil water evaporation and infiltration. These soil management practices can influence WUE by bringing about changes in net radiation, soil heat flux, sensible heat flux and photosynthetic efficiency (Hatfield et al., 2001). Increased crop residue retention is beneficial in that it provides more substrates for soil microbes, consequently increasing soil microbial biomass (Doran et al., 1998). Increased organic matter quality, favourable soil temperatures, increased soil moisture and improved soil structure result in a greater diversity in soil microbes, especially bacterial and fungal populations (Lupwayi et al., 1998, Wang et al., 2010). 3 There is limited understanding of how various moisture and tillage regimes affect WUE in chickpea and how morphological, phenological and physiological traits are associated with WUE. There is also limited understanding of which morphological, phenological and physiological traits are best used as surrogates to breed for increased WUE. Hence, the key research question for this study was: is there genetic variation for WUE in chickpea, and can surrogates be used to improve it? The overall aim of this study was to better understand how chickpea can be bred for increased WUE using morphological, phenological and physiological traits; and the effect of genotype, environment and management interactions on WUE. The specific objectives of this study were: 1. To identify genetic variation for WUE in chickpea under different moisture and tillage regimes 2. To investigate the effect of genotype by environment by management interaction on chickpea phenotypic stability 3. To understand the physiological basis of chickpea yield under water-limited conditions 4. To develop a model chickpea plant ideotype for semi-arid subtropical climates to assist plant breeding 4 2.0 LITERATURE REVIEW 2.1 Introduction Chickpea (C. arietinum L.) is among the first grain crops grown by man dating back to 7500 – 6800 BC in the Middle Eastern archeological sites (Zohary and Hopf, 2000). Its cultivation has since spread to many parts of the world due to rising interest in its high protein content, nitrogen fixing capabilities and its ability to grow in harsh conditions where other legumes cannot do well. Although it was initially a cool season crop, breeding efforts have seen its growing area expand to include the semi-arid tropics, and it has become one of the main food security crops in areas which are prone to drought. Inasmuch as chickpea is drought tolerant, it often suffers from terminal drought because it is grown on receding soil moisture in many of the cropping systems. Water is becoming increasingly scarce and development of plants that use water efficiently is one of the steps in conferring drought tolerance to plants. One of the challenges is that WUE is a complex trait. Hence, the need to explore other simple physiological traits for additive gene effects that can be used as surrogates to breed for improved WUE using both conventional and molecular techniques. 2.2 Origin and cytogenetics Substantive evidence, including unearthed seeds dating back to 5450 BC (Helbaek, 1970) and the presence of the progenitor of chickpea, Cicer reticulatum, suggest that chickpea originated in the area of southeastern Turkey adjoining Syria (Van der Maesen, 1987). From Turkey, chickpea cultivation spread in two main directions; the western province of the region, where it is grown in spring and summer, and the eastern and southern parts, where it is grown in the cool dry season (Mallikarjuna et al., 2011). De Wet et al. (1982) suggested four secondary centres of diversity, namely: the Near East region (including the Fertile Crescent), Hindustani region (current India and East Pakistan), Central Asian region (Western Pakistan, Afghanistan, Iran and south of the former Union of Soviet Socialist Republics) and the Mediterranean region (Lebanon and Palestine). Chickpea was later introduced to other parts of the world by the Portuguese and Spanish around the 1600s with kabuli types finding their way to India by the 1800s (van der Maesen, 1972). Indian immigrants imported desi chickpeas into Kenya in the 1800s (van der Maesen, 1972) and kabuli cultivars were introduced much later. Chickpea is a relatively new crop in Australia 5 with the first variety, Tyson, (a selection from C235, a northern India cultivar) released in 1978 (Berger et al., 2004). 2.3 Distribution, climate, area, production and uses of chickpea Chickpea is the third most important food legume globally after dry beans and dry peas (Parthasarathy Rao et al., 2010). It is grown mainly in the Indian sub-continent, Mediterranean basin, Australia, East Africa and the Americas. Globally, it is currently grown across 13 Mha (Foyer et al., 2016) with Asia accounting for 89% of the total area, Africa 4.6%, Oceania 2%, North America 1.6%, Latin America 1% and Europe 1%. India, which is the largest producer of chickpea in the world, accounts for 72% of total area under chickpea cultivation in Asia (which is two thirds of the global area), and is closely followed by Pakistan and Iran accounting for 11% and 7% of Asia’s chickpea cultivation area respectively (Parthasarathy Rao et al., 2010). Chickpea is primarily grown under rainfed conditions under diverse moisture and temperatures conditions with rainfall ranging from 350 mm to 600 mm annually (Malhotra and Singh, 1991). These moisture conditions vary from location to location, for example, in Australia there is variation in rainfall within the growing season among locations (Figure 1). This variation has an implication on chickpea water use efficiency. Figure 2.1: Climate data from 1981 to 2010 for chickpea growing areas in Australia adapted from Moeller and Rebetzke (2017). Closed circles indicate average maximum temperature and closed circles indicate average minimum temperatures. Bars indicate mean monthly rainfall. 1 to 12 represent month of the year where 1 = January, 12 = December. The total global chickpea production is 13 Mt with an average yield of 0.96 t ha-1 (Foyer et al., 2016). Of the main chickpea producing countries in 2012, many had low yields due to various production constraints. Ethiopia produced the highest yields of 1.7 t ha-1, followed by Australia at 1.5 t ha-1, Turkey at 1.3 t ha-1 and India at 0.9 t ha-1 (FAOSTAT, 2012). 6 Chickpea is an important legume in farming systems since it avails nitrogen to non-legume crops through biological fixation, which subsequently increases their yield and quality. Furthermore, in most cases, you don’t need to fertilise chickpea, hence it contributes to savings due to decreased use of nitrogen based fertilisers as well. Most of the positive responses expressed by cereals following legumes are primarily a result of nitrogen deposited by legumes in the previous season (Chalk, 1998). Chickpea can fix up to 140 kg Nha-1 per season which meets up to 80% of its nitrogen requirements (Saraf et al., 1998, Serraj, 2004). Unkovich et al. (2010) has also shown that chickpea can fix a range of 85 to 194 kg N ha-1. In addition, inclusion of chickpea in rotations acts as disease break and its crop residues help in maintenance of soil health and fertility through addition of organic matter, which ensures sustainability in the cropping systems. Chickpea is a very good source of protein with mature grains having a protein content of 1231%, which is among the highest in pulses (Parthasarathy Rao et al., 2010). It is also among the cheapest sources of protein (Byerlee and White, 2000), which makes it suitable for resource poor farmers, especially in developing countries. Chickpea is also a very good source of soluble and insoluble fibres, vitamins and minerals, and many other phytochemicals which are healthpromoting (Geervani, 1991). Generally, chickpea has 64% total carbohydrate, 47% starch, 6% crude fibre, 6% soluble sugars, 3% ash, and 5% fat (William and Singh, 1987). Chickpea is deficient in sulfur containing amino acids like methionine and cysteine, but rich in the essential amino acid, lysine (Sarmah et al., 2012). Chickpea is mainly used as human food and to a lesser extent as animal feed. Kabuli is mainly used as a whole grain, whereas desi can be used as whole grain or split (El-Hendawy et al.) (Sarmah et al., 2012). In some diets, chickpea seeds are eaten fresh as green vegetables, whereas in others they are parched, roasted, fried, or boiled. Chickpea can be eaten as a snack food, condiments or as stew. The seeds can be ground into flour which is used to make soup, dhal, bread, or served as a side dish (Saxena et al., 1990). Split chickpea, without its seed coat, is commonly known as dhal, which can be dried and cooked into a thick soup, or ground into flour for snacks and sweetmeats (Hulse, 1991). Gram husks, and green or dried stems and leaves are used for livestock feed. Whole seeds may be milled and used directly as feed. 7 2.4 Mode of reproduction and types of chickpea Chickpea is a self-pollinated crop and cross-pollination is rare with only 0-1% reported (Smithson et al., 1985). Self-pollination is enforced by its cleistogamous flower, whereby pollen transfer takes place before the flower opens. This may lead to a narrow genetic base having an effect on the general genetic diversity in chickpea. Chickpea is an annual diploid species divided into two types; kabuli and desi. Kabuli-types have white flowers, large, cream-coloured seeds and traditionally have been grown around the Mediterranean basin and central Asia. Desi-types have pink/purple flowers, small, dark, angular seeds and are mainly produced on the Indian subcontinent, in east Africa, central Asia and to a limited extent in the Mediterranean Basin (Cobos et al., 2007) Chickpea has an indeterminate growth habit and it can grow continuously as long as the environmental conditions especially water availability are adequate. If adequate soil moisture is present during the vegetative phase, the crops continues to be vegetative and, thus becoming a competitive sink for pod and seed formation (Khanna-Chopra and Sinha, 1990). If the crop remains vegetative for a long time, its performance at the end of the season may be affected, especially in areas which are prone to terminal drought. 2.5 Chickpea genetic resources Chickpea has three gene pools based on its crossability with the cultivated species C. arietinum (Mallikarjuna et al., 2011). The primary gene pool is comprised of cultivated species and landraces, the secondary gene pool is comprised of C. reticulatum and C. echinospermum, and the tertiary gene pool is comprised of all annual and perennial Cicer species that are not crossable with C. arietinum (Mallikarjuna et al., 2011). International Consultative Group on Agricultural Research (CGIAR) centres like the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) and the International Centre for Agricultural Research in the Dry Areas (ICARDA), and other gene repositories like the United States Department of Agriculture (USDA), maintain huge collections of cultivated chickpea comprising the primary gene pool (Mallikarjuna et al., 2011). There are 76,221 chickpea accessions conserved ex situ around the world with ICRISAT and the National Bureau of Plant Genetic Resources (NBPGR), India, holding 27% and 19% of these accessions, respectively (Foyer et al., 2016). Chickpea is a diploid with 16 chromosomes, thus; 2n = 2x = 16 (Ahmad and Hymowitz, 1993) with a genome size of 738.09 Mb. The estimated number of genes is in excess of 28,000 with 8 close to 50% of the chickpea genome comprised of transposable elements and unclassified repeats (Varshney et al., 2013a). Generally, chickpea has low genetic diversity as is the case with most legumes (Foyer et al., 2016). 2.6 Chickpea production constraints Chickpea production is affected by various production constraints; both biotic and abiotic, depending on the ecological region where it is grown. Drought resulting from water limited growing conditions is a period where soil moisture declines and can eventually lead to crop failure (Mishra and Singh, 2010). It is one of the abiotic production constraints limiting chickpea production and can either be intermittent drought occasioned by rainfall disruptions from the usual pattern of the growing season, hence leading to overall insufficient rainfall; or terminal drought occasioned by steady moisture depletion from the soil profile or less rainfall towards the end of the growing season (Canci and Toker, 2009). Drought alone causes up to 50% of chickpea production losses (Varshney et al., 2013b). However, depending on the genotype, the type of drought experienced and the environment, seed yield losses due to drought have been reported to range from 30% to 100% (Leport et al., 1999). Besides drought, other important abiotic constraints for chickpea production in Australia include high temperatures, waterlogging, boron toxicity, salinity, cold and frost. Drought and heat are also important in Kenya. Heat stress remains a major constraint, especially for cool season crops like chickpeas, and more so when they are grown in transitional and warm climatic regions (Xu and Huang, 2001). Biotic production constraints include pests like pod borers (Helicoverpa armigera, Spodoptera exigua and Helicoverpa punctigera), cutworms (Agrotis spp.), aphids (Aphis craccivora), leafminers (Liriomyza cicerina), bruchids (Callosobruchus spp.) and diseases like Ascochyta blight (Ascochyta rabiei), Fusarium wilt (Fusarium oxysporum), Phytophthora root rot (Phytophthora medicaginis), dry root rot (Rhizocotonia bataticola), collar rot (Sclerotium rolfsii) and black root rot (Fusarium solani) (Sarmah et al., 2012, Ghosh et al., 2013). The severity of pests and diseases differ from one region to another. In Australia, Ascochyta blight and Phytophthora root rot are major diseases, whereas in Kenya, Fusarium wilt and Ascochyta blight are serious diseases. In India, Fusarium wilt is the major disease affecting chickpea production but recently other diseases like dry root rot and collar rot are becoming important (Ghosh et al., 2013). Storage pests are also a major problem in India and east Africa. It is 9 estimated that 20% - 30% of stored chickpea is damaged by bruchids (Callosobruchus spp.) in South Asia (Sarmah et al., 2012). All in all, abiotic stresses cause more yield losses than biotic stresses (Sarmah et al., 2012). 2.7 Chickpea cropping systems and tillage practices The crops grown in a field over a fixed period, under a particular management system, following a specific sequence, coupled with their interaction with the farm resources, denote the cropping system. Some of the most common cropping systems include crop rotation, monocropping, intercropping and succession cropping. Sustainable cropping systems should maintain and enhance soil fertility, enhance crop growth, minimise spread of disease, weed control, enhance soil cover, reduce risk of crop failure and ensure better utilisation of resources. Highly productive and effective agricultural systems with minimal environmental damage are deemed to be important strategies for the future development of agriculture (Hanson et al., 2007). To attain these goals, there is a need to develop production systems which are diverse and intensely managed (Kirschenmann, 2007). These, coupled with ecologically based management principles employed in dynamic cropping systems in farmlands, leads to sustainability. This will ensure agricultural production based on a strategy of annual crop sequencing which optimises production, resource conservation and economic returns (Hanson et al., 2007). Crop sequencing generally increases the WUE of a cropping system (Merrill et al., 2007) since the plants make maximum use of the available soil moisture. Legumes perform well in crop sequences and offer a great opportunity to sustain increased productivity because of their ability to adapt to different cropping patterns (Jeyabal and Kuppuswamy, 2001) and fix atmospheric nitrogen. They also help reduce soil erosion (Giller and Cadisch, 1995) and suppress weeds (Exner and Cruse, 1993) if included as an intercrop in a cropping system. Chickpea fits into various cropping systems (which vary from region to region). They include sole crop, mixed or intercropped, however, it is mainly intercropped with barley, linseed, mustard, maize, peas, safflower, sorghum and coffee among others (Berrada et al., 2007). Chickpea is grown in rotation following wheat, barley or rice (van der Maessen L.J.G., 1972). Wheat-chickpea sequential cropping has been successfully used in Australia, Ethiopia and Spain, and rice-chickpea sequences in Nepal, Bangladesh and Eastern India (Garrido and Lopez-Bellido, 2001, López-Bellido et al., 1998, Zewdu, 2002, Harris et al., 2005, Felton et 10 al., 1998). A decline in arable land has led to the integration of chickpea into sequential cropping systems where it is grown under irrigation or receding soil moisture (Berrada et al., 2007). Sequential cropping is recommended in chickpea cultivation since growing chickpea in the same field repeatedly is highly discouraged due to the risk of diseases like Ascochyta (P. rabiei) (Berrada et al., 2007). Crop rotation helps lower pest and disease pressure in cropping systems by causing a break of suitable host for the pest or disease organisms. Cyst nematodes (Heterodera cicero) can be controlled by rotating chickpea with non-leguminous crops (Ahlwat and Shivakumar, 2003). Intercropping reduces the incidence of pests and diseases in chickpea compared with a chickpea monocrop (Berrada et al., 2007). The incidence of crown rot in wheat is always lower when wheat is grown following chickpea compared with wheat grown consecutively in the same field (Felton et al., 1998). Rotations including pulse crops in wheat fields, especially in no till systems, minimise the damage caused by cereal root diseases as well as increase the population and activity of beneficial soil microorganisms (Krupinsky et al., 2002). Wheat yields increased by 810 kg ha-1 and 1360 kg ha-1 in 1989 and 1990, respectively, when wheat was grown after chickpea as opposed to after wheat at Warra in Queensland, Australia. Similarly, wheat grain protein content increased from 9.4% to 10.7% (Hossain et al., 1996). Wheat shoot dry biomass and nitrogen increased by an average of 0.85 tha-1 and 19.2 kg N ha-1, respectively, when wheat was grown after chickpea in northern New South Wales, Australia (Felton et al., 1998). Pulse crops, including chickpea, pose a soil erosion threat since they produce lower crop residues with lower carbon to nitrogen (C/N) ratios when compared to grain cereals (Berrada et al., 2007). In this regard, it is imperative to grow chickpea alongside cereal crops that produce large amounts of residues to ensure residue retention especially in conservation tillage. 2.8 Conservation agriculture and its implication on chickpea cultivation Conservation agriculture is founded on three main principles; minimal soil disturbance, soil cover with crop residues and crop rotation. This broad system of management helps in the improvement of soil fertility, disease and weed control (Verhulst et al., 2010). The principle of minimal soil disturbance encompasses reduced tillage systems whereby at least 30% of the soil surface is covered by crop residue between harvesting and the next planting (Fowler and Rockstrom, 2001). Zero till is a form of conservation tillage which ensures no more than 2025% of the soil surface is disturbed with seeding performed using narrow slits into untilled soils (Sayre and Govaerts, 2012). Zero till has been successfully implemented in over 96 Mha 11 of rainfed production systems in the USA, Brazil, Argentina, Canada and Australia (Derpsch, 2005). The use of conservation tillage has increased since its inception in the 1960s, primarily due to its ability to lower farm resource requirements, soil erosion control and soil moisture conservation (Verhulst et al., 2010). It also increases soil aggregate stability (Chan and Mead, 1988, Li et al., 2007), improves soil structure (Page et al., 2013), increases water storage, especially in the semi-arid regions (Marley and Littler, 1989, Felton et al., 1995, Radford et al., 1995) due to increased infiltration rates and reduced evaporation. Increased infiltration rates occur due to the continuity of macropores created by plant roots from the previous crop and soil fauna, particularly earthworms. Macropores act as channels that help transport water into the lower soil horizons. Crop residues on the soil surface and increased aggregate stability prevent the formation of surface seals, which normally impede infiltration in soils (McGarry et al., 2000). The crop residues also lower soil temperature and reduce soil surface wind speeds, consequently reducing water loss through evapotranspiration (Jones et al., 1994, Hatfield et al., 2001). Due to reduced levels of soil disturbance in conservation tillage, soil bulk density increases (Li et al., 2007). Consequently, soil porosity decreases (Mielke et al., 1986), and, if coupled with an increase in soil moisture, it leads to decreased air permeability and a reduced number of airfilled pores. This causes an increase in anaerobic processes such as denitrification in wet periods (Linn and Doran, 1984). . Several researchers (Muñoz-Romero et al., 2012, Jan et al., 2012, Gan et al., 2010) have looked into the productivity of chickpea under no till regimes. Jan et al. (2012) reported that chickpea planted under conventional tillage yielded more than the no till treatment. This may have been a result of higher plant density observed in conventional tillage systems as opposed to no till. In addition, Muñoz-Romero et al. (2012) reported that chickpea root length was higher in conventional tillage than no till systems. However, there was no significant difference in root biomass in the two tillage systems. Chickpea root length, root biomass and root nitrogen decreased with increasing soil depth under both conventional and no till systems (MuñozRomero et al., 2012). In wet years, root distribution was highest in the superficial soil layer as opposed to drier years where there was a higher distribution in the deeper soil layers (MuñozRomero et al., 2012). Chickpea roots have the ability to grow more than 1 m deep in the semiarid regions, so they can scavenge for water in the deeper soil horizons (Gan et al., 2010). Chickpea water use has been demonstrated to be higher in tilled-fallow systems than no-till 12 systems and increased with increasing soil depth in Saskatchewan, Canada (Gan et al., 2010). Despite all these studies, the results have not been conclusive. 2.9 Drought resistance mechanisms in plants Plants respond to water deficit conditions in various ways depending on the duration and intensity of the deficit, and the stage of plant development. The main defence mechanisms against drought include escape, avoidance and tolerance. Drought escape can either be through early flowering or early vigour. In most cases, early flowering genotypes mature early in the season to ‘escape’ terminal drought, especially in the semi-arid areas where this is the norm. An example of a very early flowering chickpea type includes ICCV 96029, and an early maturity type includes ICCV 2 (Gaur et al., 2008). The disadvantage with early maturing genotypes is that they tend to be smaller in stature and consequently have a lower photosynthetic area, and in most cases, have lower yield potential (Blum, 1988). That is why it is important to match genotypes with environment to take advantage of the maximum growing duration with least amount of stress for optimum yields. Early vigour is an equally important drought escape mechanism in grain legumes (Thomson and Siddique, 1997). Early plant vigour should also be matched with the right environment. Dehydration avoidance is the ability of the plant to maintain turgor in its tissues and cells under water deficit conditions by maintaining water uptake and reducing water loss. Long roots allow access to water deep in the subsoil. ICC 4958 has large roots which develop quickly to rapidly extract water from the subsoil (Toker et al., 2007). Other mechanisms involved in maintaining water uptake and reducing water loss include osmotic adjustment, which maintains stomatal conductance and photosynthesis and, in effect, delays leaf senescence and death (Toker et al., 2007). Leaf characteristics like glandular droplets consisting of organic acids such as succinic, malic, citric and oxalic (Toker et al., 2004) help in lowering the leaf temperature, thereby protecting the plant from drought (Lauter and Munns, 1986). Drought tolerance is the ability of the plant to tolerate water deficits with low tissue water potential and maintain metabolic function at low leaf water status. Some of the mechanisms that can be exploited to confer drought tolerance include an ability to remobilise stem reserves to fill grains, maintenance of cell membrane stability and accumulation of abscisic acid during water stress conditions. Others mechanisms include proline accumulation, presence of polyamines, brassinosteroids, jasmonates, phosphatidic and salicylic acids (Toker et al., 2007). 13 2.10 Effects of water deficit on chickpea growth and development Chickpea is indeterminate crop, which has the habit of continuous vegetative growth if there is no water limitation. However, it can quickly change from vegetative to reproductive phase. The challenge is normally that chickpea is grown under receding soil moisture and there is a high probability that by the time the plant changes from vegetative to reproductive, there is insufficient soil moisture for the reproductive phase, hence leading to seed yield loss. Generally, water deficit causes a reduction in seed yield due to various factors as shown in previous studies. Seed yield, pods per plant and average seed size were all higher on primary branches than on secondary branches when moisture stress was induced at the early podding stage (Leport et al., 2006). Chickpeas exposed to terminal drought under field conditions had a shorter seed filling duration and seed filling rate and thus, had smaller final seed size (Davies et al., 1999). However, early transient water deficit increased the rate of seed filling and final seed size at maturity compared with well-watered controls (Fang et al., 2011). Moisture stress at the early podding stage reduces the number of seeds per pod to predominantly one in kabuli, whereas desi-types are not affected (Leport et al., 2006). Fewer seeds per pod and smaller seed size result in a decrease in seed yield (Leport et al., 1999, Fang et al., 2010), although at late podding, moisture stress does not reduce the seed size (Leport et al., 2006). Similar data were reported by Davies et al. (1999) who reported the average seed size of chickpea genotypes Tyson, ICCV 88201 and Kaniva were reduced by 19, 23 and 34%, respectively, under field conditions. Reduction in seed yield under the early transient water deficit was lower compared with terminal drought yield losses (Fang et al., 2011), probably because of chickpea’s indeterminate nature and ability to recover. Water stress reduced the total plant dry mass (Behboudian et al., 2001), particularly in kabuli types where there was greater pod number and yield reduction, than in desi-types (Leport et al., 1999). Terminal drought reduced seed yield by 58-95% compared with irrigated controls (Leport et al., 2006). Chickpea is most sensitive to water deficit during the flowering and early podding stages (Khanna-Chopra and Sinha, 1987). A study by Fang et al. (2011) showed that early transient water deficit reduced flower production by almost 50% and increased flower abortion compared with the well-watered controls in two chickpea cultivars, Rupali, (a desi-type) and Almaz (a kabuli-type). Terminal drought caused a 33-63% reduction in flowering, and an increase of 37-56% in flower abortion and 54-73% in pod abortion (Fang et al., 2010). The rate of flower abortion was higher for flowers on secondary branches than on primary branches, 14 and for late produced flowers than those produced earlier in the season (Fang et al., 2010). When pre-dawn leaf water potential was below -1.2 MPa, flower abortion occurred as a result of low pollen viability and failure of pollen tubes to grow down the style resulting in no fertilisation. Water deficit impaired the function of both pollen and the stigma/style – pollen germination was low with fewer pollen tubes reaching the ovary (Fang et al., 2010), which is characteristic of angiosperms grown under stressed environments (Porch and Jahn, 2001). Pod abortion is more sensitive to water stress in kabuli than desi types, and kabuli tends to yield less than desi under similar conditions (Leport et al., 2006). Pod formation is greatly reduced after moisture stress is induced (Behboudian et al., 2001) although if induced at late flowering, it has a minimal effect on pod production. Pod abortion is higher in pods borne on secondary branches compared with those borne on primary branches regardless of when the stress is induced (Leport et al., 2006). Leport et al. (2006) reported that the total number of pods was reduced by 66-75% in plants exposed to early podding moisture stress compared with wellwatered controls 2.11 Chickpea physiological responses to water deficit Many plant morphological and physiological processes are affected by water deficit conditions (Toker and Cagirgan, 1998). These include reduced water content and water potential, stomatal closure, turgor loss and cell enlargement and plant growth (Rahbarian et al., 2011). Lower soil water potential in drying soils slows plant growth (Ohashi et al., 2000), reduces photosynthesis (Gren and Quist, 1985), affects hormonal balance (Munns and Cramer, 1996), reduces cell enlargement (Nonami et al., 1997) and slows cell division as a result of reduced cyclin-dependent kinase activity (Schuppler et al., 1998b). Relative water content is a very good indicator of a plant’s response to water deficit as it indicates the hydration status of leaves (Barrs and Weatherley, 1962). When chickpea plants were exposed to water deficit conditions in pots, the relative water content of the leaves decreased (Krouma, 2010). Matos et al. (2010) reported similar responses for two chickpea genotypes whereby the relative water content was lower under water stress compared with the well-watered control. Water deficit can reduce the relative water content of chickpea genotypes at seedling, flowering or podding stages (Rahbarian et al., 2011). Leaf water potential, an indicator of plant water status has been reported to be lower in drought tolerant chickpea than drought sensitive genotypes under drought stress conditions (Rahbarian 15 et al., 2011). Similar data have been reported by Siddique et al. (2000) in wheat, although there are also contrasting reports for wheat and other crop species. The water potential measured in the leaves of three chickpea genotypes, Chetoui, Kesseb and Andoun, under water deficit decreased by 1.5, 1.6 and 2.1 fold, respectively over a 21 day period (Krouma, 2010). In general, Andoun was more drought tolerant than the other genotypes. Among the first signals of water stress is stomatal closure, which consequently slows photosynthesis as a result of limited carbon dioxide availability to the mesophyll (Chaves, 1991). Several factors including leaf water deficit (Hsiao, 1973), soil water deficit and leaf to air water vapour pressure deficit (Schulze, 1986) can cause stomatal closure. Water stress reduced stomatal conductance by 28-70% compared with well-watered controls in three chickpea genotypes (Krouma, 2010). Transpiration and transpiration efficiency was higher in well-watered chickpea plots compared with water stressed plots (Singh and Sri Rama, 1989). Krouma (2010) reported a decrease in transpiration of between 27-61% in three chickpea genotypes under water stress. Photosynthesis is generally inhibited under water stress conditions due to stomatal closure and various factors including the imbalance between light capture and utilisation (Foyer and Noctor, 2000), a decrease in the internal carbon dioxide concentration, inhibition of ribulose1,5-biphosphate carboxylase/oxygenase enzyme activity and ATP synthesis (Rahbarian et al., 2011). In three water stressed chickpea genotypes, net photosynthesis was decreased by 3351% compared with the well-watered controls (Krouma, 2010). This decrease later translated into yield penalties, which was related to the strong positive correlation (0.965) between biomass and photosynthesis in water stressed plants compared with a moderate positive correlation (0.50) in the well-watered plants. A decrease in photosynthesis at the flowering and podding stages for chickpea genotypes exposed to water stress compared with well-watered controls was related to a decrease in internal CO2 concentration (Rahbarian et al. (2011). The down-regulation of photosynthesis causes an energy imbalance in photosystem II, which results in photoinhibition (Pastenes et al., 2005). Rahbarian et al. (2011) reported a decrease in photosystem II efficiency in chickpea genotypes under water stress. Photosystem efficiency (Fv/Fm) helps in the detection of any damage to photosystem II and its probable inhibition. Water stress affects photosystem efficiency and thus, decreases the electron transport rate and the effective quantum yield of photosystem II (Ahmed et al., 2002). Stomatal conductance, net photosynthetic rate and photosynthetic capacity were reduced in chickpea under water stress 16 conditions but recovered after rehydration (Matos et al., 2010). The recovery upon rehydration shows that, inasmuch as water stress slows down photosynthesis, it does not damage the photosynthetic apparatus (Zanella et al., 2004), though this largely depends on the level of water stress. Water stress can reduce chlorophyll a and b levels, which in turn alters their light harvesting capabilities (Farooq et al., 2009). Sayed (2003) pointed out that water stress decreases chlorophyll a/b binding proteins and, in effect, impairs the synthesis of chlorophyll a/b, thus leading to a reduction in light harvesting pigment protein associated with photosystem II. The thylakoid membrane emits chlorophyll fluorescence and it can be used as a proxy for photosynthetic reaction in photosystem II (Ahmed et al., 2002). Damage to the light reaction systems in photosynthetic apparatus as a result of water stress can be detected by analysing chlorophyll fluorescence and photosynthetic efficiency (Rahbarian et al., 2011). Membrane stabilisation is important under water stress conditions and it can be achieved through changes in lipid composition or preservation of membrane lipids (Thi et al., 1990). Under water stress conditions, cell membranes experience dysfunction, causing increased levels of ion permeability and leakage (Sayar et al., 2008). Changes in membrane stability are thus identified by measuring electrolyte leakage from leaf discs in solution (Blum and Ebercon, 1981). An increase in electrical conductivity of the solution indicates increased membrane damage. Rahbarian et al. (2011) and Matos et al. (2010) reported reduced membrane stability in chickpea genotypes under water stress compared with well-watered controls. Moreover, membrane injury was higher when the relative water content was ≤40% compared with when the relative water content was 55-50% (Matos et al., 2010), demonstrating that chickpea cell membranes become less stable with increasing severity of the water stress. 2.12 Water use efficiency and associated breeding efforts Water use efficiency has various meanings depending on the discipline of study (Passioura, 2006) and can also be interpreted at various scales including farm, field, plant and plant part levels (Morison et al., 2008). Water use efficiency in agriculture can be considered at the whole plant level (ratio of total dry matter produced to total water used), economic yield (ratio of crop grain per unit area to transpiration), and at the leaf level (ratio of instantaneous carbon dioxide assimilation rate to transpiration rate) during the growing season (Ali and Talukder, 2008). For the purposes of this thesis, future reference of water use efficiency from Chapter 4 onwards 17 will be to water use efficiency in agriculture (ratio of total dry matter produced to total water used). At the crop level, water loss is a result of the difference in water vapour concentration between the crop canopy and the atmosphere, and it is least during the cool humid months of the year. At the leaf level, the rates of CO2 assimilation (A) and transpiration (T) are a product of stomatal conductance (gs) and also a concentration gradient between the inside and outside of the leaf for CO2 and water vapour, respectively (Condon et al., 2002). Theoretically, intrinsic water use efficiency (WT = A/gs) can be improved by lowering the ratio between intracellular to atmospheric CO2 concentration (Ci/Ca), although trade-offs are likely to occur (Condon et al., 2002). However, breeding efforts have been made to select for lower Ci/Ca values that are reflected as low stomatal conductance values, high photosynthetic capacity or a combination of both (Farquhar et al., 1989). There is substantial genetic variation for Ci/Ca determined through carbon isotope discrimination (CID), which is large enough to cause variation in A/T, and consequently, WUE for dry matter production (Farquhar and Richards, 1984). Rebetzke et al. (2002) showed that CID is a highly heritable trait for wheat, which can be manipulated through plant breeding. Thus, increasing the intrinsic WUE has been an attractive crop breeding target over the years (Fischer, 1981). By exploiting genetic variation associated with intrinsic earliness and response to photoperiod, breeders have developed genotypes that can grow in times of the year when the evaporative demand is low which in turn raises the ratio of A/T and increases yield (Condon et al., 2004). Chickpea has a slow initial growth rate and low photosynthetic rate, hence low WUE (Singh et al., 1987). As the crop progresses, WUE at the biomass level increases from the vegetative stage and peaks at full flowering and thereafter, decreases again towards maturity (Singh et al., 1987). WUE varies depending on the environment grown. WUE for grain ranged from 7 to 20 kg ha-1 mm-1 in a tropical climate compared to 10 to 13 kg ha-1 mm-1 in a sub-humid temperate climate (Table 2.1). 18 Table 2.1: Water use efficiency (ratio of total dry matter produced to total water used) of chickpea genotypes across different environments Site Climate type Plant part WUE (kg Reference(s) ha-1 mm-1) Thohoyandou, Tropical Grain production Limpopo (summer crop) Above 7 – 20.9 ground 12 – 41.1 Province, South biomass Africa production Canterbury, Sub-humid Grain production 10 - 13 New Zealand temperate Above ground 22 - 29 Ogola and Thangwana, 2013 Anwar et al., 2003 biomass production Warra, Humid Sub- Grain production 5.9 Queensland, tropical Above ground 14.2 Australia Dalal et al., 1997 dry matter Windridge, Humid Sub- North Star, tropical Total dry matter 29.2 Grain 8.8 Herridge et al., 1995 NSW, Australia Glenhoma, Humid Sub- North Star, tropical Grain 5.8 8.7 NSW, Australia Tel Hadya, Mediterranean Above ground Northern Syria climate biomass Zhang et al., 2000 production Grain 3.2 2.13 Breeding for increased water use efficiency using surrogates Target trait based breeding has gained primacy in recent years as opposed to breeding for increased yields generally. Plant breeders are selecting for physiological traits that are simple to work with as surrogates for traits that have been traditionally difficult to select for. Drought tolerance is a very complex trait and WUE, which is one of the major components of drought 19 adaptation, is complex too. Hence, it is of prime importance to improve related traits that give an additive gene effect to increase WUE and drought tolerance. Some traits associated with WUE have been identified, which include CID, where low CID implies higher transpiration efficiency resulting from low stomatal conductance or high rates of CO2 assimilation (Condon et al., 2002, Farquhar et al., 1989). Delayed leaf senescence in sorghum is related to higher WUE through greater biomass accumulation post-anthesis (Borrell et al., 2014) and spike photosynthesis improves WUE in cereals probably due to re-fixation of respiratory CO2 and better maintenance of water status through osmotic adjustment (Araus and Tapia, 1987). These traits can be used as surrogates for WUE (Reynolds and Tuberosa, 2008) 2.14 Phenotyping target physiological traits in chickpea Over the last century, breeders have made progress in improving drought tolerance by selecting constitutive traits that affect dehydration avoidance rather than drought responsive traits because of fewer yield penalties (Blum, 2006). Target traits in water-limited environments should be correlated with yield and should have higher heritability than yield (Monneveux and Ribaut, 2006). Phenotyping these traits should also be non-destructive, accurate, cheap and inexpensive (Tuberosa, 2011). The phenotypic performance needs to be associated with genotypic data to understand the genetic basis of these complex traits (Montes et al., 2007). For phenotyping to be successful and relevant, environmental characterisation (Tuberosa, 2011, Chenu et al., 2011) is vital so that genotype by environment interactions can be exploited (Trethowan, 2014). Phenotyping of large plant populations for various traits in the field can be labour intensive and expensive. However, the emergence of high-throughput phenotyping platforms such as near infra-red spectroscopy and multi-spectral reflectance make it possible to phenotype some simple traits in large populations in multi-locations (Montes et al., 2007). Unmanned aerial platforms such as polycopters mounted with cameras further increase the data capture and resolution, hence, increasing the output of the system (Araus and Cairns, 2014). Chickpea phenotyping for drought tolerance has focused on selection for early maturity to avoid drought, and root traits to confer improved WUE (Upadhyaya et al., 2011). Phenotyping for WUE in chickpea has mainly been conducted using gravimetric methods in a pot culture (Upadhyaya et al., 2011); however, these methods do not generally correlate well with field conditions. 20 Near infrared spectroscopy has been used to capture differences in dry matter, starch and crude protein of maize (Montes et al., 2007). Spectral reflectance allows monitoring of various dynamic complex traits using high temporal resolution without destroying the plant (Montes et al., 2007). It can be used to measure canopy architecture and nitrogen concentration (Montes et al., 2007). Other measurements can be made on individual plants including plant photosynthesis pigment composition and plant water status. (Peñuelas and Filella, 1998). Examples of some of the data that can be captured for chickpea breeding programs include morphological, phenological, physiological data (Figure 2.2). High transpiration efficiency (Kashiwagi et al., 2006c, Turner et al., 2001) Active water use strategy until flowering (Kashiwagi et al., 2013) Deep and more profuse roots (Kashiwagi et al., 2005) Cooler canopies at mid reproductive stage (Purushukothaman et al., 2015) High rate of partitioning / sink activity (Krishnamurthy et al., 2013, Krishnamurthy et al., 1999) Strong root system (Krishnamurthy et al., 2004, Kashiwagi et al., 2006a) Figure 2.2: Some target traits for chickpea physiological breeding 2.14.1 Canopy temperature Canopy temperature has been used as an indirect indicator of crop water status in cereals since water deficit results in partial stomatal closure, thus reducing transpiration and in effect causing sunlit leaves to become warmer than the ambient temperature (Jackson et al., 1977). Since transpiration has a cooling effect on canopies, cooler plant canopies can indicate higher transpiration rates, which is also a function of available soil water. Other factors that affect canopy temperature include morphological traits like leaf angle, canopy architecture, waxy deposits or other compounds that reflect heat (Pietragalla, 2012, Tuberosa, 2012), agronomic 21 traits like plant density and tillage (Yang et al., 2014) and atmospheric conditions like incident radiation, wind and relative humidity (Mariano et al., 2012). Under water limited conditions, cooler canopy temperatures are related to the capacity of plants to extract soil water from deep in the subsoil, whereas under well-watered conditions sink strength and photosynthetic capacity are more important (Pietragalla, 2012). The hand held canopy temperature gun is a simple and rapid method of determining canopy temperatures, however, in very large fields it may be limiting. Thermal imagery systems are more amenable to high throughput phenotyping for canopy temperature in large experiments (Kashiwagi et al., 2008) and these can be achieved by mounting the thermal imagery systems (e.g. cameras) on unmanned aerial platforms like drones, polycopters and airplanes. Canopy temperature is quite sensitive to environmental conditions and caution should be taken while taking the measurements. Good results are achieved when the conditions are ideal for high vapour pressure deficit (VPD), in conditions of warm air, generally above 15°C and relative humidity of less than 60% with clear sunny skies and low wind speeds (Pietragalla, 2012). 2.14.2 Plant vigour and plant green biomass Over the years, remote sensing imagery has gained popularity because it is not limited by sampling interval or geostatistical interpolation (Moran et al., 1997), does not involve destructive sampling, and is amenable to high throughput. The premise for using optical remote sensing for crop assessment is that crop canopy multispectral reflectance and temperature is associated with photosynthesis and evaporation in which leaf area index (LAI) and crop development stages are central (Bauer, 1985). Several indices have been developed which are used to analyse aerial imagery (Shanahan et al., 2001) including the normalised difference vegetation index (NDVI). The NDVI links reflectance in the red region and the near infra-red (NIR) to vegetation parameters such as canopy cover, leaf area index and the concentration of total chlorophyll (Shanahan et al., 2001). Korobov and Railyan (1993) concluded that the NIR and red areas of the spectrum correlated highly with plant parameters such as plant height, plant density and percent plant cover. Initially, the NDVI was used for estimating green biomass (Tucker, 1979), however it has been subsequently used to assess crop health (Douglas Ramsey et al., 1995, Teillet, 1992). The use of NDVI in breeding has been made possible by the development of inexpensive equipment that is simple to use, affordable and accurate. 22 2.14.3 Photosynthetically active radiation (PAR) The photosynthetic active radiation spectrum (PAR), which makes up 50% of the total global radiation (Bonhomme, 2000), lies in the wavelength 400 – 700 nm (Zhang et al., 2008). The crop canopy absorbs PAR, referred to as intercepted photosynthetically active radiation (IPAR) which is intercepted light used for photosynthesis and eventually producing plant biomass (Johnson et al., 2010). The radiation intercepted during the growing period is determined by the canopy radiation extinction coefficient (k) and is influenced by leaf orientation and the green leaf area (Thomson and Siddique, 1997). Research has shown that lower k values are associated with narrow and erect leaves compared to plant genotypes with more horizontal leaf arrangements (Kiniry et al., 2005). Lower k values allow more light to penetrate the canopy and illuminate more leaf area in conditions of low light intensity, thus increasing carbon exchange rates, and consequently, radiation use efficiency (Kiniry et al., 2005). The fraction of intercepted photosynthetically active radiation can be used to estimate the leaf area index (LAI) through its relationship with the plant canopy (Johnson et al., 2010). This provides an easy and non-destructive way of estimating the leaf area index. IPAR can be accurately determined using a ceptometer, though care should be taken to avoid confounding factors such as the soil albedo, row spacing and lack of canopy uniformity (Andrade et al., 2002). 2.14.4 Chlorophyll content There is a close relationship between chlorophyll concentration, leaf nitrogen content and crop yield (Cartelat et al., 2005). This relationship arises because the majority of leaf nitrogen is usually contained in chlorophyll (Cartelat et al., 2005). Since chlorophyll absorbs PAR, which aids in photosynthesis, it indicates the strength of the internal leaf apparatus during photosynthesis (Li et al., 2006). Leaf chlorophyll content can be determined by extraction with organic solvents including acetone (Liu et al., 2008) and methanol (Cenkci et al., 2010) and subsequent quantification using a spectrometer; however this method is expensive and time consuming (Jangpromma et al., 2010). A higher throughput non-destructive method is the SPAD chlorophyll meter that allows rapid and inexpensive assessment of leaf greenness (Ahmed, 2011). SPAD measures leaf absorbance in the red (650 nm) and infrared (940 nm) regions (Markwell et al., 1995), and gives readings that have been correlated with chlorophyll content under different moisture regimes in many crops (Jangpromma et al., 2010). 23 2.15.5 Root traits Plants extract water from the soil through the roots and the spatial distribution of the root system influences water and nutrient intake capacity (Lynch, 1995). Dense root systems are more efficient at extracting water from the top soil horizon whereas deeper rooting systems are better at extracting water from the lower soil horizons. These contrasting traits are important influencers on yield under water deficit conditions during the reproductive stage in many crops (Ludlow and Muchow, 1990). Kashiwagi et al. (2006a) showed that root architecture of chickpea affects transpiration by influencing soil moisture use and subsequent harvest index under terminal drought. However, the heritability of these root characteristics will determine their utility in plant breeding. Varshney et al. (2014) reported genetic variation for both root length density and root depth in chickpea and found heritabilities ranging from medium to low. Root hydraulic conductivity impacts the amount and efficiency of water uptake by the plant and is determined by the anatomy and morphology of the roots and their aquaporin activity (Bramley et al., 2009). In legumes, root hydraulic conductance is influenced by the total root length since water is absorbed along the full root (Bramley et al., 2009). Root phenotyping is difficult and for this reason the literature on chickpea is not extensive. However, Kashiwagi et al. (2006a) and Zaman-Allah et al. (2011) used polyvinyl chloride (PVC) cylinders (lysimeters) to grow chickpeas for assessment. The soil was subsequently washed off sampled plants to measure total rooting depth. Image analysis software was then used to estimate the root length at various sections of the lysimeters and divided by the specific volume of that section to determine the root length density. With these advances in root phenotyping, many plants can be assessed. 2.15 Chickpea ideotype development Plant breeders empirically select for yield in their breeding program. This selection is based on variation created through hybridisation or by introduction of various genotypes with varying responses to the trait of interest. Inasmuch as this method has led to yield increases over the years, it still poses a challenge in that little is known about the morphological, physiological and biochemical determinants of yield. Furthermore, yield is highly affected by the environment due to its polygenic nature and thus, affects the repeatability of the results over different seasons (Johnson and Geadelmann, 1989). 24 The ideotype approach is an alternative strategy to empirical breeding where a deliberate attempt is made to understand the factors that influence yield formation under different abiotic stresses. Donald (1968) defined an ideotype as a biological plant model that behaves in a known manner when exposed to a distinct environment. The idea behind the ideotype was to consolidate several important plant traits into one genotype, which would be ideal for growing in a specified environment. The definition of the plant type assists plant breeders to have more clear cut objectives (Rasmusson, 1987, Rasmusson, 1991) and thus, a blueprint for pyramiding traits (Mock and Pearce, 1975). This makes ideotype breeding more analytical than the traditional empirical selection and breeding methods used in the past. Key steps in ideotype breeding include the identification of the target population of environments (Mock and Pearce, 1975, Trethowan, 2014). The ideotype should perform optimally in the defined target environment. The second step entails identification of the physiological and morphological traits that contribute to yield either directly or indirectly. These traits should have genetic diversity and be highly heritable to be incorporated into an ideotype breeding program (Rasmusson, 1987). The identification of both morphological and physiological traits can be done through physiological breeding (Figure 2.2). 25 Step 1: Trait identification based on association with yield • Select experimental conditions to match target environment • Selection of traits with high variability and heritability • Selection of experimental material with similar phenology but contrasting genetic potential • Selection of an experimental site similar to target environment • Design and implement proper protocols for data capture • Measure the association of the trait with yield Step 2: Quantifying heritability and gains from selecion • Develop populations from contrasting parents with respect to traits of interest • Measure heritability and gains obtained from selection • Integrate selected traits into breeding program Figure 2.3: Schematic illustration of physiological breeding. Adapted from Reynolds et al. (2001). 2.16 Crop modelling and ideotype design From a modelling perspective, crop ideotype is a set of defined crop parameters that drive growth and development in defined environmental conditions (Rotter et al., 2015). This entails the use of high quality, long-term data, for model calibration and the generation of accurate simulation results (Rotter et al., 2015). Data from multiple sites over many years can be produced without running actual field trials, thus creating a powerful tool for ideotype design and testing in silico (Semenov and Stratonovitch, 2013). This in turn saves a lot of time and money that would have been used to test the genotypes in a wide set of environments. Agricultural Production Systems sIMulator (APSIM) is an important software package used in agricultural crop modelling. It simulates cropping systems using climate, soil, management and crop genetic coefficients to predict the economic yield of a crop species (Keating et al., 2003). The APSIM model mainly employs the supply and demand concept of important plant growth 26 resources (light, water, nitrogen and carbon) to create a plant phenotype (Hammer et al., 2001). This is mainly based on the input parameters which are therefore used to give the output of the plant trait being modelled. 2.17 Conclusion Tillage systems may increase WUE, but the results to date are inconclusive. Hence, more research needs to be conducted to elucidate how tillage systems affect WUE in chickpea. Furthermore, the effect of genotype by environment by management in chickpea has not been extensively studied, hence the need to delve further in this area of study. A few surrogate traits, including carbon isotope discrimination, have been identified that can be used to select for drought tolerant genotypes. However, previous studies mostly identified one or two surrogates and in single environments. There is a need to identify multiple surrogates in different environments, to develop a chickpea ideotype that can perform well in a target environment because previous attempts to develop chickpea ideotypes have used only a few traits resulting in poor crop ideotypes. The proposed ideotype traits are either labeled as low, medium or high and this makes it difficult for the plant breeder to know what is high or low without figures. There should be an attempt to guide plant breeders with a quantitative trait range with values, e.g., low should be X1 to X2 within the available genepool. 27 CHAPTER 3: GENERAL MATERIALS AND METHODS 3.1 Introduction This chapter covers the general materials and methods for the field experiment. Each chapter has more specific materials and methods including formulas. 3.2 Experimental site The field experiment was carried out at the University of Sydney’s Plant Breeding Institute at Narrabri (30.275616°S and 149.803547°E) in 2014 and 2015. This site has a summer dominant rainfall and in winter the rainfall is not sufficient for a successful crop. Hence, crops grown during winter, including chickpea, tend to experience terminal drought. On average, the longterm annual rainfall is 662 mm distributed throughout the year with a peak in December and January. The long-term mean annual maximum and minimum temperatures are 26.5°C and 11.7°C, respectively, with the coldest month being July. The soil at the site is characterised by deep Vertosols, which are black clays that shrink and expand with changes in soil moisture. The planting window for chickpea in Narrabri is from the second week of May to the second week of June according to the annual winter crop sowing guide produced by New South Wales Department of Primary Industries (DPI). 3.3 Experimental design The field experiment was planted under no till and till systems with each having irrigation splits such that there was no till, +/- irrigation and till, +/- irrigation (Figure 3.2) using an alpha lattice design replicated twice. The irrigated side which received two supplementary irrigations was considered to be the well-watered treatment and the rainfed side was considered to be the water stress treatment. There were 30 chickpea entries (25 desi and 5 kabuli), in addition, five genotypes were selected based on their phenotypic similarity, then mixed to form six mixture entries (Table 3.1) totalling 36 entries in the experiment. The genotypes were sourced from Pulse Breeding Australia (PBA) in Tamworth except for the ICCV lines which were sourced from ICRISAT India. Additionally, Sal, Sim, Lyle, Lyons, Austin and Doolin were sourced from the University of Sydney germplasm store. The PBA lines were selected because they are grown widely and also some of them have drought tolerance to some extent. The ICCV lines have been tested in ICRISAT India and there was a need to further test them for water use efficiency. The preceding crop in the experimental area was wheat in both seasons planted in rotation, such that chickpea is not planted in the same field where the previous crop was chickpea. 28 Zero tillage → Tillage → Rainfed Irrigated ↓ ↓ Rep 1 Rep 1 Rep 2 Rep 2 Rep 1 Rep 1 Rep 2 Rep 2 Figure 3.1: Narrabri experimental field layout in 2014 and 2015 for chickpea water use efficiency experiments Table 3.1: List of chickpea genotypes for water use efficiency experiments at Narrabri No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Name AMETHYST FLIPPER GENESIS KALKEE HOWZAT KYABRA PBA HATTRICK PBA SLASHER PBA STRIKER SONALI TYSON YORKER GENESIS 079 GENESIS 090 ICCV 96853 ICCV 98801 ICCV 98813 ICCV 98816 ICCV 98818 Type Desi Desi Kabuli Desi Desi Desi Desi Desi Desi Desi Desi Kabuli Kabuli Desi Desi Desi Desi Desi No. 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 Name JIMBOUR JIMBOUR#1 FLIP 079C ICCV 05308 AUSTIN DOOLIN HOWARD LYLE LYONS SAL SIM THOMAS Mix 1 (Yorker/Jimbour) Mix 2 (Howzat/Flipper) Mix 3 (Flipper/Jimbour) Mix 4 (Yorker/Howzat) Mix 5 (Howzat/Jimbour) Mix 6 (Howzat/ 98813) Type Desi Desi Kabuli Kabuli Desi Desi Desi Desi Desi Desi Desi Desi Desi Desi Desi Desi Desi Desi 3.4 Field experiment sowing Sowing was carried out using a six-row planter with 30 cm inter-row spacing in 2014 and a four-row planter with 50 cm inter-row spacing in 2015 resulting into 4 m by 2 m plots in both years. Plant population was maintained at 25 plants m-2 for both years. The date of sowing was 28 May, 2014 and 11 June, 2015. The four-row planter was used in 2015 because the stubble in the no till area was high and it was difficult for the six-row planter to cut through it. Seeds 29 were dressed using P-Pickle® T (360 g/L thiram and 200 g/L thiabendazole) at 2 mL in 8 L of water with 1 kg of solution used per kg of seed, and later with Apron® XL 350 ES (350 g/L metalaxyl M) at 0.75 mL in 9.25 L of water with 10 mL of solution used per kg of seed, in both years. These fungicides were used to give protection against fungal diseases during the early stages (normally up to six weeks) of crop development. The seeds were inoculated with chickpea group N rhizobia (Nodulaid®) using the slurry method in 2014 and as a solution (water + inoculum) injected into the soil using a tank mounted on the planter in 2015 at the recommended label rates. 3.5 Field agronomic practices Pre-emergence spray Terbyne® 750WG (750 g/kg terbuthylazine) was applied at 1 kg ha-1 and Balance® 750WG (750 g kg-1 isoxaflutole) at 100 g ha-1 for weed control in the field. During the cropping season, any weeds present were pulled out manually in the experimental area. Prophylactic sprays using Ridomil Gold® at 2.5 kg ha-1 were applied at flowering and midpodding to protect the crop against phytophthora root rot (Phytophthora medicaginis) in both years. Unite® 720 (720 g/L chlorothalonil) was applied at the rate of 500 mL ha-1 for the control of Ascochyta blight (A. rabiei) at early flowering in 2014 and at early flowering and early podding in 2015. Insect pests, mainly caterpillars (Helicoverpa armigera) and aphids, were controlled using Karate Zeon® (250 g/L lambda-cyhalothrin) at the rate of 36 mL/ha in 2015. 3.6 Data parameters Several parameters were measured during the growing season and some post-harvest traits were also measured (Table 3.2). The main foci were agronomic, morphological, phenology and physiological data. 3.7 Field irrigation Two irrigations were applied in both seasons using a lateral moving sprinkler irrigation system. In 2014, 35 mm was applied at flowering and early podding, whereas in 2015, 36 mm was applied at flowering and 26 mm at late podding/early maturity stage (represented as inverted arrows in Figure 3.3c and 3.3d). 3.8 Weather data Weather data was collected from the nearest weather station at Narrabri Airport in 2014 and from the Managed Environment Facility weather station in a nearby field in 2015. Data for rainfall, temperature (maximum and minimum), radiation and evaporation were recorded from time of sowing to harvesting. During the growing season, total rainfall was considered as the 30 rainfall received during the active plant growing period, thus from sowing time to when the plants in the experiment reached 75% maturity. 3.9 Data analysis Data were analysed using Genstat® edition 18 following the methods described in the individual chapters. 31 CHAPTER 4: WATER USE, WATER USE EFFICIENCY AND YIELD VARIATION IN CHICKPEA GENOTYPES 4.1 Introduction Chickpea is mainly grown on stored soil water in areas where it is cultivated (Kashiwagi et al., 2005). As such, the crop has to strike a balance in water use to ensure that there is enough soil moisture towards the end of the growing season and at the same time to have extracted enough water to sustain yield. Legumes mainly have either a conservative water use strategy, where water is used sparingly, or a profligate water use strategy where the water use is more liberal (Bacelar et al., 2012). These water use strategies determine the survival of the crop, especially under water limiting conditions since their survival is dependent on moisture availability at the reproductive stage (Kato et al., 2008). This was emphasised by Zaman-Allah et al. (2011) who posited that chickpea genotypes that are drought susceptible used more water at the vegetative stage whereas the drought tolerant genotypes used less water at the vegetative stage and more water at the reproductive stage. Deep and profuse rooting systems are very important in accessing soil water from deep down the soil horizon (Kashiwagi et al., 2006b) and can give chickpea plants a reprieve under water limited conditions. Supplementary irrigation during the flowering and pod filling stages has been shown to increase seed yield as well (Silim and Saxena, 1993). Water use efficiency is an important trait in crops grown under stored soil water as well as under irrigation (Blum, 2005). Water use efficiency has various definitions depending on the level and measurement scale, but for the purpose of this chapter it will be defined as the ratio of grain yield to water used (Condon et al., 2004). There have been reports that there is genetic variation for WUE in various crops (Farquhar and Richards, 1984). This may give plant breeders an opportunity to exploit this trait in improving crop yields under water limiting conditions, especially under stored soil water. Improvement of WUE requires a multifaceted strategy (Wang et al., 2002) which includes breeding and management (Condon et al., 2004). Some of the management practices that increase WUE include crop sequencing since it ensures maximum use of available soil water (Merrill et al., 2007). Tillage and no till systems also affect water use efficiency in different ways and it is imperative to understand their effect in order to incorporate them in the management options. Early flowering in chickpea is used as a drought escape mechanism and helps the crop avoid seed yield losses as a result of terminal drought. This ensures that the plant will produce some grain even though there will be a yield penalty due to the low moisture conditions at the end of the growing season. 32 This in effect increases agronomic water use efficiency where seed yield is considered per unit amount of water used. Indeterminate flowering may cause the crop to delay in flowering and end up losing yield at the end of the growing season if moisture is inadequate. Little attention has been paid to the pattern of water use in legumes and the relationship between water used and seed yield (Zhang et al., 2000b). Despite evaluation of WUE in chickpea in various studies, little has been achieved since these studies were focused on single factors affecting WUE. This causes variability of data from different studies due to failure of integration of various factors (Gan et al., 2010). Studies conducted by Angadi et al. (2008) showed WUE of 6.8 kg ha-1 mm-1 for chickpea grown on the Canadian Prairies whereas McKenzie et al. (2006) reported 15.8 kg ha-1 mm-1 for the same location. It is therefore important to incorporate more factors to get more reproducible data. Soil factors (tillage and fertility) also play a key role in minimising variation in the data. However, there have been limited studies on the effects of tillage systems on chickpea production. A few preliminary studies showed the benefits of no-till management were primarily due to soil moisture conservation and availability in the growing season (Rathore et al., 1998). There is also very little knowledge about how soil moisture status and WUE for chickpea is affected by cropping systems (Gan et al., 2010). The hypotheses to be tested in this chapter include:  whether there is genetic variation for WUE in chickpea and  whether no till systems conserve more soil water and increase WUE in chickpea, relative to conventional cultivation. This chapter aims to: i) discover whether there are differences in water use and WUE of chickpea genotypes, ii) evaluate the effect of tillage and irrigation management options on WUE, iii) establish the relationships among water use, WUE and yield and iv) establish heritability estimates for WUE under different management options. 4.2 Materials and methods Thirty-six entries were grown for two years (2014 and 2015) at the IA Watson Plant Breeding Institute in Narrabri, under no till and till conditions with irrigation (well-watered/non-stress) and without irrigation (water stressed) as described under general Materials and Methods in Chapter 3. Soil moisture in the control plots was monitored on a fortnightly basis using a neutron probe moisture meter, CPN® 503DR Hydroprobe (Figure 4.1) from sowing on a 33 fortnightly basis until harvesting. A total of 16 (in 2014) and 32 (in 2015) aluminium neutron probe access tubes (Figure 4.2) were inserted immediately after sowing up to a depth of 150 cm spread across the whole experiment in all the control plots. The control plots in 2014 included PBA Hattrick and Tyson whereas in 2015 they included PBA Hattrick, Tyson, Amethyst and Sonali genotypes. Measurements were taken at 10, 20, 30, 40, 60, 80, 100, 120 and 134 cm in every tube. Figure 4.1: Neutron probe moisture meter Figure 4.2: Neutron probe access tube The neutron probe moisture meter was set to take counts for 16 seconds and then data recording started. The data were converted to volumetric water content (θ) in millimetres using the equation (4.1) below from a soil calibration exercise in the Managed Environment Facility in Narrabri. θ = (C – 7863)/182.9 (4.1) Where, θ is the volumetric water content in millimetres and C is the neutron counts. The soil water balance method (Equation 4.2) was used to estimate water use which was estimated to be equivalent to evapotranspiration from planting to physiological maturity as documented by Anwar et al. (1999). WU = Et = (P + I) - ΔSWC - Ro – D (4.2) Where WU is water use, Et is evapotranspiration, P is precipitation, I is irrigation, ΔSWC is change in soil water content from time of measurement 1 to 2 at a depth of 0-134 cm, Ro is run-off and D is drainage. Run-off was assumed to be zero since there was no major rain event to necessitate a run-off and the fields were effectively level. Drainage was set at zero because soil drained upper limit was not reached during the cropping season. Total water use was 34 considered as the initial water at the beginning of the season less the remaining water at the end of the season and also taking into consideration irrigation and precipitation. The total seed yield was obtained by harvesting and threshing all the plants in the plot, weighing them and then converted to yield in kilograms per hectare. Water use efficiency for grain production was calculated as the total grain produced divided by the total water used and expressed as kg ha-1 mm-1. This was done for the control genotypes which had neutron probe access tubes fitted. For the rest of the genotypes, WUE was calculated by taking the average water use from the control genotypes in each environment type and divided by the total seed yield for each genotype in that environment type. Each year was considered separately. Broad sense heritability was calculated as the ratio of genotypic to phenotypic variance (Equation 4.3) (Knapp and Bridges, 1987) H2 = σ2g/(σ2g + σ2ge/e + σ2e/re) (4.3) Where σ2g is the genotypic variance, σ2ge is the genotype by year variance component and σ2e is the error term variance. “e” and “r” represents the year and replication, respectively. The genotypic coefficient of variation (GCV), phenotypic coefficient of variation (PCV) and environment coefficient of variation (ECV) were calculated as the ratio of the standard deviation of each variation to the trait mean as shown in Equations 4.4 to 4.6. GCV = (√σ2g)/ X̅, (4.4) PCV = (√σ2p)/ X̅, (4.5) ECV = (√σ2e)/ X̅ (4.6) Where, σ2g, σ2p, σ2e and X̅ are the genotypic variance, phenotypic variance, error variance and the trait mean, respectively. Genetic advance (GA) as a percentage of the mean was calculated using Equation 4.7. GA = ((K*√σ2p* H2)/ X̅)*100 (4.7) Where √σ2p is the phenotypic standard deviation, H2 is the heritability and K is the selection differential at 5% selection intensity (2.06) 35 Data were analysed using Genstat® edition 18 for all the measured traits using linear mixed models in Restricted Maximum Likelihoods (REML) (Patterson and Thompson, 1971) to estimate variance components. Tillage, moisture, year and genotypes were fitted in the fixed model whereas the range and rows were in the random model (Fixed model: Tillage X Moisture X Year X Genotype; Random model: Range.Row). To get least significant differences, the model was changed to tillage, moisture and genotypes in the fixed model and range and row were nested within the years in the random model (Fixed model: Tillage X Moisture X Genotype; Random model: year/(Range.Row). Data on water use was analysed by considering Tillage X Moisture X Genotype in the fixed model. Tillage, moisture and year had two factor levels whereas genotypes had 36 factor levels. Data means, standard error of the means, coefficient of variation and least significant difference were tabulated. 4.3 Results 4.3.1 Precipitation and temperature The total rainfall received in 2014 was 132.5 mm and 156.5 mm in 2015, respectively (Figure 4.3c and 4.3d). However, rainfall distribution during the flowering phase was better in 2014 than 2015. In general temperatures were higher in 2014 than in 2015 especially during the flowering and podding phase (Figure 4.3a and 4.3b). 36 40 a 30 30 20 20 10 10 b o T em p era tu re ( C ) 40 0 0 50 100 100 150 -1 0 -1 0 40 R a in fa ll (m m ) 50 150 50 c d 40 30 30 20 20 10 10 0 0 0 50 100 150 0 50 100 150 D a y s a f t e r s o w in g Figure 4.3: Daily maximum and minimum temperatures for 2014 (a) and 2015 (b) and rainfall for 2014 (c) and 2015 (d) at the water use efficiency experimental field site: Narrabri. Inverted arrows on the rainfall graphs show the amount and timing of irrigation water applied. 4.3.2 Seed yield The combined two year seed yield average was 1722 kg ha-1 under irrigation and 1478 kg ha-1 under rainfed conditions (Figure 4.4a) with a range of 1223 to 2074 kg ha-1 under irrigation, and 1172 to 1849 kg ha-1 under rainfed conditions. Supplementary irrigation increased seed yield by 16.5% (Figure 4.4a). The average yield under no-till conditions was 1658 kg ha-1 which was 7.4% higher than the average yield under traditional tillage (Figure 4.4b). There was high seasonal variation with average yields in 2014 (1894 kg ha-1) being significantly higher than in 2015 (1307 kg ha-1) (Figure 4.4c). 37 M o is tu r e T illa g e S ea so n 2500 b Y ie ld in k g h a -1 a c 2000 1500 1000 500 0 I r r ig a t e d R a in f e d N o - T ill T ill 2014 2015 Figure 4.4: Mean chickpea grain yields under irrigation and rainfed conditions (a), no till and till systems (b), and different seasons (c). Error bars in the graph represent standard errors of the mean. The highest yielding environment was the no-till, irrigated plot in 2014 with an average yield of 2148 kg ha-1, whereas the lowest yield was from the standard till, rainfed plot in 2015 (Table 4.1). Sonali was the highest yielding genotype in both irrigated and rainfed no-till systems in 2014 with a yield of 2680 and 2210 kg ha-1, respectively. The lowest yielding genotype in both environments was ICCV 05308 with a yield of 1446 and 1107 kg ha-1 in irrigated and rainfed no-till systems, respectively. PBA Slasher had the highest yield in 2015 under no-till plus irrigation, whereas ICCV 96853 yielded the highest under no-till rainfed conditions. Genesis 079 had the highest yield under till plus irrigation in 2014, whereas Sonali had the highest yield under similar conditions in 2015. Sonali had the highest yield in 2014 under till and rainfed conditions, whereas PBA Slasher had the highest yield under the same conditions in 2015. However, the performance of these genotypes was not stable due to strong interactions among season, moisture regime and genotype. Table 4.1: Mean chickpea seed yield (kg ha-1) under different management and experimental conditions. No-Till No-Till Till Till Irrigated Rainfed Irrigated Rainfed Genotype 2014 2015 2014 2015 2014 2015 2014 2015 AMETHYST 2336 1317 1360 1438 2227 1401 1725 972 AUSTIN 2019 1556 1409 1427 1785 1287 1673 1324 DOOLIN 2228 1570 1657 1439 1888 1306 1517 1150 FLIP 079C 2227 1044 1724 962 2100 1195 2092 1177 38 FLIPPER 1966 1344 1604 1053 1850 1485 1570 1063 GENESIS 079 2287 952 1756 900 2474 1301 1571 1183 GENESIS 090 1884 1476 1841 1330 2057 1335 1464 854 GENESIS KALKEE 1813 1012 1574 876 1623 863 1445 793 HOWARD 2013 1454 1705 1273 1848 1346 1629 1050 HOWZAT 2193 1370 1968 1471 2144 1262 1753 1256 ICCV 05308 1446 1293 1107 1338 1210 943 1650 1137 ICCV 96853 2557 1616 1990 1695 2340 1567 1890 1284 ICCV 98801 2266 1458 1668 1563 1968 1264 1887 1270 ICCV 98813 1544 1380 1290 1257 1654 1294 1291 1037 ICCV 98816 1986 1378 1428 1268 1615 1192 1702 1137 ICCV 98818 1769 1436 1637 1340 1693 1137 1535 1033 JIMBOUR 2428 1658 1796 1551 2075 1335 1854 1370 JIMBOUR#1 2049 1704 1703 1312 1916 1489 1775 1226 KYABRA 2171 1328 1904 1208 2076 1197 1933 1002 LYLE 2182 1506 1716 1244 1951 1270 1576 1212 LYONS 2128 1431 1616 1499 2047 1392 1579 1205 Mix 1 (Yorker/Jimbour) 2185 1493 1874 1284 1964 1292 1722 895 Mix 2 (Howzat/Flipper) 2170 1468 1592 1398 1987 1454 1668 1153 Mix 3 (Flipper/Jimbour) 2317 1455 1907 1388 1998 1308 1750 1120 Mix 4 (Yorker/Howzat) 2252 1360 1623 1468 2216 1363 1629 1037 Mix 5 (Howzat/Jimbour) 2368 1489 1983 1358 2207 1337 1823 1254 Mix 6 (Howzat/ 98813) 2295 1407 1769 1685 1932 1568 1792 1289 PBA HATTRICK 2340 1503 2000 1468 2154 1422 1778 1283 PBA SLASHER 2419 1859 1941 1326 2381 1635 1968 1388 PBA STRIKER 2257 1171 2063 1388 2232 1251 1716 1194 SAL 1850 1505 1608 1264 1806 1352 1513 1006 SIM 2050 1428 1648 1336 2065 1356 1737 1090 SONALI 2680 1492 2210 1666 2318 1675 2197 1324 THOMAS 2136 1627 1575 1251 1926 1164 1682 1242 TYSON 2460 1353 1965 1258 2285 1134 1839 1177 YORKER 2057 1434 1543 1296 1952 1220 1707 1024 Mean yield (kg ha-1) 2148 1426 1715 1341 1999 1316 1712 1145 39 Pooled SE 175 175 175 175 175 175 175 175 Pooled LSD @ 5% 348 348 348 348 348 348 348 348 4.3.3 Seed yield variation and interaction under different tillage and moisture regimes There was a significant difference (P<0.001) between the no-till and till systems with the notill plots having higher yields than the till plots (Table 4.2). Irrigated plots had higher yields than the rainfed treatments. There was a significant difference (P<0.001) in the yield of genotypes as well as their performance across the two seasons (2014 and 2015). There was a significant two-way interaction between year and tillage, year and moisture, and year and genotype, as well as a significant (P<0.001) three-way interaction between tillage, moisture and year (Table 4.2), indicating that seed yield largely depends on seasonal weather conditions. The year effect was a key driver of the interaction and also explained a lot of the variation in the data. Table 4.2: Wald statistic for main effects (tillage, moisture, genotype, season) and their interaction on chickpea seed yield Fixed term Wald statistic Tillage 60.15 *** Moisture 273.12 *** Genotype 401.57 *** Year 1641.74 *** Tillage.Moisture 0.85 Tillage.Genotype 35.16 Moisture.Genotype 42.51 Tillage.Year 6.94 * Moisture.Year 64.69 *** Genotype.Year 156.19 *** Tillage.Moisture.Genotype 41.12 Tillage.Moisture.Year 15.67 Tillage.Genotype.Year 23.91 Moisture.Genotype.Year 37.9 Tillage.Moisture.Genotype.Year 30.46 *** *=P<0.05, ** = P<0.01 and ***=P<0.001 40 4.3.4 Water use Soil water measurements at the start of the experiment showed that soil volumetric water content ranged from 354 – 453 mm in 2014, and 351 – 492 mm in 2015. There was higher average water use by plants under no till than till experimental plots in both years (Figure 4.5a). Water use under no till was 355 mm in 2014 and 301 mm in 2015. Plants under tillage used 332 mm water in 2014 and 296 mm in 2015 (Figures 4.5a and 4.5c, respectively). In general, more water was used in 2014 than 2015. Crops under irrigation used 359 mm in 2014 and 316 mm in 2015. Under rainfed conditions, crops used 328 mm in 2014 and 280 mm in 2015 (Figures 4.5b and 4.5d, respectively). Total water use in the 2014 growing season was 319 mm under the no-till rainfed regime, 278 mm in the till rainfed plots, 364 mm in no-till irrigated and 355 mm in the till irrigated plots. The total water use for 2015 was lower than in 2014 with no-till rainfed using 283 mm, till rainfed using 309 mm, no till irrigated using 364 mm and till irrigated using 355 mm of water. T illa g e M o is tu r e 400 a b 300 2014 W a te r u se in m m 200 100 400 c d 300 2015 200 100 0 N o T ill T ill I r r ig a t e d R a in f e d Figure 4.5: Mean chickpea water use under no till and till systems in 2014 (a) and 2015 (c), and under irrigation and rainfed conditions in 2014 (b) and 2015 seasons (d) 41 Plants accessed moisture in the top soil layer (up to 30 cm) in the early growing days and reached a peak extraction around the flowering period. In 2014, plants had deeper roots which extracted soil moisture from as far as 100 cm below the soil surface (Figure 4.6a), compared with a depth of 60 cm from flowering onwards in 2015 (Figure 4.6b). There was a sharp decline in soil water from flowering onwards in both 2014 and 2015 (Figures 4.6a and 4.6b). Soil moisture levels at 120 cm and 134 cm did not change over time meaning plants roots did not reach that far and deep water drainage was not occurring. b V o lu m e tr ic w a te r c o n te n t (m m ) a 80 80 20 60 30 60 40 40 60 40 80 20 20 0 0 100 120 0 50 100 D a y s a fte r s o w in g 150 134 0 50 100 150 D a y s a fte r s o w in g Figure 4.6: Mean volumetric water content (denoting root water access) at various soil depths during the growing season in 2014 (a) and 2015 (b). The legend on the right hand side shows soil depth in centimetres from the ground surface up to 134 cm deep in the soil profile. 4.4.5 Water use efficiency under different tillage and moisture regimes (individual analysis) Water use efficiency under no-till irrigated conditions in 2014 ranged from 4.0 to 7.4 kg ha-1 mm-1 with a mean of 5.9 kg ha-1 mm-1. In total, 21 genotypes performed above the trial mean (Table 4.3). Under the same tillage and moisture conditions in 2015, WUE ranged from 3.0 to 5.8 kg ha-1 mm-1 with a mean of 4.5 kg ha-1 mm-1 with 22 genotypes above the trial mean. Under no-till and rainfed conditions, WUE ranged from 3.2 to 6.4 kg ha-1 mm-1 and 3.1 to 6.3 kg ha-1 mm-1 with means of 5.0 kg ha-1 mm-1 and 4.7 kg ha-1 mm-1 in 2014 and 2015, respectively (Table 4.3). The mean performance for WUE under till and irrigated conditions was 5.6 and 4.2 kg ha1 mm-1 in 2014 and 2015, respectively (Table 4.3). The range in these conditions were 3.4 to 7.0 kg ha-1 mm-1 in 2014, and 2.7 to 5.2 kg ha-1 mm-1 in 2015, with 17 genotypes having better WUE than the trial mean in each year (Table 4.3). The range of WUE under till and rainfed conditions was 4.2 to 7.1 kg ha-1 mm-1 in 2014, and 2.8 to 5.2 kg ha-1 mm-1, with a mean of 5.5 42 kg ha-1 m-1 and 4.1 kg ha-1 mm-1 in 2014 and 2015, respectively. The number of genotypes that outperformed the trial mean in the same environments was 18 in 2014, and 19 in 2015 (Table 4.3). Of the eight environments tested, Sonali had the highest water use efficiency in five environments (no-till irrigated 2014, no-till rainfed 2014 and 2015, till rainfed 2014 and 2015), PBA Slasher in two (no-till irrigated 2015 and till irrigated 2015) and Genesis 079 in one environment (till irrigated 2014). The lowest WUEs were recorded in three environments for ICCV 05308 (no-till irrigated 2014, no-till rainfed 2014 and till-irrigated 2014), three for Genesis Kalkee (no-till rainfed 2015, till-irrigated 2015 and till rainfed 2015), with Genesis 079 and ICCV 98813 in one environment each (no-till irrigated 2015 and till rainfed 2014, respectively). Table 4.3: Mean chickpea WUE among genotypes under different tillage, irrigation and seasonal conditions No-Till No-Till Till Till Irrigated Rainfed Irrigated Rainfed 2014 2015 2014 2015 2014 2015 2014 2015 AMETHYST 6.4 4.2 3.9 5.0 6.3 4.3 5.6 3.3 AUSTIN 5.5 4.9 4.1 5.0 5.0 4.1 5.4 4.8 DOOLIN 6.1 4.9 4.8 5.1 5.3 4.2 4.9 4.1 FLIP 079C 6.1 3.3 5.0 3.4 5.9 3.8 6.7 4.2 FLIPPER 5.4 4.2 4.6 3.7 5.2 4.7 5.1 3.8 GENESIS 079 6.3 3.0 5.1 3.2 7.0 4.2 5.1 4.2 GENESIS 090 5.2 4.6 5.3 4.7 5.8 4.3 4.7 3.1 GENESIS KALKEE 5.0 3.2 4.6 3.1 4.6 2.7 4.7 2.8 HOWARD 5.5 4.6 4.9 4.5 5.2 4.3 5.3 3.8 HOWZAT 6.0 4.3 5.7 5.2 6.0 4.0 5.7 4.5 ICCV 05308 4.0 4.0 3.2 4.7 3.4 3.0 5.3 4.1 ICCV 96853 7.0 5.1 5.8 6.0 6.6 5.0 6.1 4.6 ICCV 98801 6.2 4.6 4.8 5.5 5.6 4.0 6.1 4.6 ICCV 98813 4.2 4.3 3.7 4.4 4.7 4.1 4.2 3.7 ICCV 98816 5.5 4.3 4.1 4.5 4.6 3.8 5.5 4.1 ICCV 98818 4.9 4.5 4.7 4.7 4.8 3.6 5.0 3.7 JIMBOUR 6.7 5.2 5.2 5.5 5.9 4.3 6.0 4.9 JIMBOUR#1 5.6 5.4 4.9 4.6 5.4 4.7 5.7 4.4 Genotype 43 KYABRA 6.0 4.2 5.5 4.3 5.9 3.8 6.2 3.6 LYLE 6.0 4.7 5.0 4.4 5.5 4.1 5.1 4.4 LYONS 5.8 4.5 4.7 5.3 5.8 4.4 5.1 4.3 Mix 1 (Yorker/Jimbour) 6.0 4.7 5.4 4.5 5.5 4.1 5.6 3.2 Mix 2 (Howzat/Flipper) 6.0 4.6 4.6 5.0 5.6 4.6 5.4 4.1 Mix 3 (Flipper/Jimbour) 6.4 4.6 5.5 4.9 5.6 4.2 5.6 4.0 Mix 4 (Yorker/Howzat) 6.2 4.3 4.7 5.2 6.3 4.3 5.3 3.7 Mix 5 (Howzat/Jimbour) 6.5 4.7 5.7 4.8 6.2 4.3 5.9 4.5 Mix 6 (Howzat/ 98813) 6.3 4.4 5.1 5.9 5.4 5.0 5.8 4.6 PBA HATTRICK 6.5 4.5 5.8 5.0 6.4 4.9 5.7 4.7 PBA SLASHER 6.6 5.8 5.6 4.7 6.7 5.2 6.3 5.0 PBA STRIKER 6.2 3.7 6.0 4.9 6.3 4.0 5.5 4.3 SAL 5.1 4.7 4.6 4.5 5.1 4.3 4.9 3.6 SIM 5.6 4.5 4.8 4.7 5.8 4.3 5.6 3.9 SONALI 7.4 4.8 6.4 6.3 6.5 4.9 7.1 5.2 THOMAS 5.9 5.1 4.5 4.4 5.4 3.7 5.4 4.5 TYSON 6.7 4.2 5.6 4.5 6.1 3.8 6.0 4.0 YORKER 5.6 4.5 4.5 4.6 5.5 3.9 5.5 3.7 Mean 5.9 4.5 5.0 4.7 5.6 4.2 5.5 4.1 Pooled SE 0.66 0.58 0.66 0.58 0.66 0.58 0.66 0.58 Pooled LSD @ 5% 1.06 1.16 1.06 1.16 1.06 1.16 1.06 1.16 Where SE is the pooled standard error and LSD is the pooled least significant difference at 95% confidence interval. 4.3.6 Water use efficiency under different tillage and moisture regimes (combined analysis) Combined analysis for WUE in the two years (2014 and 2015) was done for the four environments (no-till irrigated, no-till rainfed, till irrigated and till rainfed) and the range for no-till irrigated was 4.0 to 6.2 kg ha-1 mm-1 with a mean of 5.19 kg ha-1 mm-1 (Table 4.4). PBA Slasher had the highest WUE and ICCV 05308 had the lowest. Under the no-till rainfed system, the range for WUE was 3.8 to 6.4 kg ha-1 mm-1 with a mean of 5.85 kg ha-1 mm-1. Sonali had the highest WUE efficiency in this environment and Genesis Kalkee had the lowest. WUE ranged from 3.2 to 6.0 kg ha-1 mm-1 in the till and irrigated environment with a mean of 4.92 44 kg ha-1 mm-1 with PBA Slasher and ICCV 05308 having the highest and lowest WUE respectively. Under the no-till rainfed conditions, Sonali had the highest WUE and Genesis Kalkee the lowest with the range of 3.7 to 6.1 kg ha-1 mm-1 and a mean of 4.82 kg ha-1 mm-1 (Table 4.4). Table 4.4: Mean chickpea WUE for combined analysis in 2014 and 2015*. No-Till No-Till Till Till Irrigated Rainfed Irrigated Rainfed AMETHYST 5.3 4.5 5.3 4.4 AUSTIN 5.2 4.6 4.6 5.1 DOOLIN 5.5 4.9 4.7 4.5 FLIP 079C 4.7 4.2 4.9 5.5 FLIPPER 4.8 4.2 5.0 4.4 GENESIS 079 4.6 4.1 5.6 4.7 GENESIS 090 4.9 5.0 5.0 3.9 GENESIS KALKEE 4.1 3.8 3.7 3.7 HOWARD 5.0 4.7 4.8 4.5 HOWZAT 5.2 5.4 5.0 5.1 ICCV 05308 4.0 4.0 3.2 4.7 ICCV 96853 6.0 5.9 5.8 5.3 ICCV 98801 5.4 5.2 4.8 5.3 ICCV 98813 4.3 4.1 4.4 3.9 ICCV 98816 4.9 4.3 4.2 4.8 ICCV 98818 4.7 4.7 4.2 4.3 JIMBOUR 5.9 5.3 5.1 5.4 JIMBOUR#1 5.5 4.8 5.1 5.1 KYABRA 5.1 4.9 4.8 4.9 LYLE 5.4 4.7 4.8 4.7 LYONS 5.2 5.0 5.1 4.7 Mix 1 (Yorker/Jimbour) 5.3 5.0 4.8 4.4 Mix 2 (Howzat/Flipper) 5.3 4.8 5.1 4.8 Mix 3 (Flipper/Jimbour) 5.5 5.2 4.9 4.8 Mix 4 (Yorker/Howzat) 5.2 4.9 5.3 4.5 Mix 5 (Howzat/Jimbour) 5.6 5.3 5.2 5.2 Genotype 45 Mix 6 (Howzat/ 98813) 5.4 5.5 5.2 5.2 PBA HATTRICK 5.5 5.4 5.7 5.2 PBA SLASHER 6.2 5.2 6.0 5.7 PBA STRIKER 4.9 5.4 5.1 4.9 SAL 4.9 4.6 4.7 4.3 SIM 5.1 4.7 5.1 4.8 SONALI 6.1 6.4 5.7 6.1 THOMAS 5.5 4.5 4.6 4.9 TYSON 5.4 5.1 5.0 5.0 YORKER 5.1 4.5 4.7 4.6 Mean 5.19 4.85 4.92 4.82 Pooled SE 0.56 0.56 0.56 0.56 Pooled LSD @ 5% 0.94 0.94 0.94 0.94 * LSD for the combined analysis (2014 and 2015) derived by including year in the random model On average, no-till plants had a higher WUE (5.02 kg ha-1 mm-1) than plants grown in the till plots (4.87 kg ha-1 mm-1) and the irrigated plants (5.05 kg ha-1 mm-1) had higher WUE than the rainfed (4.84 kg ha-1 mm-1) (Figure 4.7-1 and 4.7-2, respectively). The highest environment mean for WUE was recorded in no-till irrigated (IRN) plots (5.19 kg ha-1 mm-1) followed by till irrigated (IRC) (4.92 kg ha-1 mm-1) (Figure 4.6-4). This was followed by no-till rainfed (RFN) (4.85 kg ha-1 mm-1) and then till rainfed (RFC) (4.82 kg ha-1 mm-1) which had the lowest WUE among the environments (Figure 4.7-4). Water use efficiency was higher in 2014 with a mean of 5.51 kg ha-1 mm-1 compared with 2015 with a mean of 4.38 kg ha-1 mm-1 (Figure 4.73). T illa g e ) -1 1 T illa g e x M o s itu r e 3 2 b 5 4 a a a RFN IR C RFC 4 W U E (k g h a -1 mm S ea so n M o is tu r e 6 3 2 1 0 N o T ill T ill I r r ig a t e d R a in f e d 2014 2015 IR N Figure 4.7: Chickpea WUE under different tillage (1), moisture (2), season (3) and tillage by 46 moisture interaction (4). IRN, irrigated no till; RFN, rainfed no till; IRC, irrigated till; and RFC, rainfed till. Different letters in the right panel (no.4) indicate significant difference at P<0.05. 4.3.7 Genetic variation for water use and WUE under different tillage and moisture regimes Water use was similar (P>0.05) among the genotypes evaluated in both years (Table 4.5). However, there was a significant difference (P<0.01 in 2014 and P<0.001 in 2015, respectively) in water use between water regimes with more water being used under irrigation compared with rainfed conditions. There was a significant tillage effect (P<0.01) on water use in 2014 but not in 2015. There was no significant interaction (P>0.05) for water use among all the treatments. Table 4.5: Variation for chickpea water use among genotypes under different tillage and moisture regimes Source of variation Mean sum of squares 2014 2015 Combined Tillage 2096.3** 166.1 1365.9 Moisture 3971.5** 10000.0*** 13932.1*** Genotype 815.7 105.2 2128.2 Tillage.Moisture 696.4 0.7 211.4 Tillage.Genotype 113.9 956.6 1275.1 Moisture.Genotype 549.3 895.2 876.4 Tillage.Moisture.Genotype 312.6 588.7 253.8 Residual 171.8 514.6 895.8 *=P<0.05, ** = P<0.01 and ***=P<0.001 There were highly significant differences (P<0.001) for the genotype, moisture, genotype and year main effects in both years (2014 and 2015) for all the environments except moisture main effect in 2015 (Table 4.6). Analysis for each individual year showed two way interactions between tillage and moisture for both years. Much of the variation in both years was accounted for by genotypic differences. Combined analysis for 2014 and 2015 showed that all the main effects were highly significant (P<0.001) except for tillage which was significant at P<0.01. In the combined analysis, significant two way interactions were observed in tillage by moisture, tillage by year, moisture by year, and genotype by year, whereas three way interactions were 47 observed for tillage by moisture by year. Much of the variation under the combined analysis was accounted for by variation in the year followed by genotypic differences. Table 4.6: Components of variation in chickpea WUE in 2014 and 2015 Source of variation Wald statistic 2014 2015 Combined Tillage 6.15* 44.25*** 10.28** Moisture 69.07*** 1.88 20.83*** Genotype 345.46*** 173.41*** 374.41*** Year 605.14*** Tillage.Moisture 40.18*** 6.60* 5.71* Tillage.Genotype 18.64 38.54 34.78 Moisture.Genotype 37.64 36.77 41.20 Tillage.Year 44.02*** Moisture.Year 46.98*** Year.Genotype 130.82*** Tillage.Moisture.Genotype 40.18 28.12 39.90 Tillage.Moisture.Year 40.28*** Tillage.Year.Genotype 24.68 Moisture.Year.Genotype 33.82 Tillage.Moisture.Year.Genotype 26.21 *=P<0.05, ** = P<0.01 and ***=P<0.001 4.3.8 Water use, water use efficiencyand yield relationships under rainfed and irrigated conditions There was a moderate positive correlation between water use and yield under rainfed conditions at r2 = 0.46 (Figure 4.8a) and a high positive correlation between water use and yield under irrigated conditions r2 = 0.75 (Figure 4.8c). The more the genotypes used water under irrigated conditions, the higher the yield. This was not the case always under rainfed conditions, where some genotypes would use a lot of water but the yield would still remain low. Water use efficiency was highly and positively associated with yield. However, the association was stronger with r2 = 0.96 under irrigated conditions (Figure 4.8d) compared with rainfed conditions with r2 = 0.78 (Figure 4.8b). Genotypes with high water use efficiency had the highest yield while those with low water use efficiency had the lowest yields. 48 W a te r u s e e ffic ie n c y W a ter u se a 2000 1500 1000 R 2 = 0 .4 6 R 2 = 0 .7 8 Y = 8.086*X - 883.9 Y = 337.8*X - 208.5 c d 2500 2000 Ir r ig a te d Y ie ld i n k g / h a b R a in f e d Y ie ld i n k g / h a 2500 1500 1000 R 2 = 0 .7 5 R Y = 14.63*X - 3126 500 250 300 350 400 W a te r u se in m m 2 = 0 .9 6 Y = 440.3*X - 541 4 5 6 7 8 W U E in k g / h a /m m Figure 4.8: Relationships between water use, water use efficiency and yield. a) Water use vs yield under rainfed conditions, b) water use efficiency vs yield under rainfed conditions, c) water use vsyield under irrigated conditions d) water use efficiency vs yield under irrigated conditions 4.3.9 Heritability and genetic advance of WUE Heritability was low under no-till systems with no-till irrigated, and no-till rainfed plants having heritability estimates of 36.4% and 43.3%, respectively (Table 4.7). High heritability was found under till systems where till rainfed had the highest heritability of 73% and till irrigated 71.3%. Genetic advance was higher under the till system compared with no-till which had moderate genetic advance. 49 Table 4.7: Heritability estimates and genetic advance for chickpea genotypes under different tillage and moisture regimes Parameter Heritability (%) Genetic advance (GAM) Till Irrigated 71.3 20.4 No-Till Irrigated 36.4 10.4 Till Rainfed 73.0 21.4 No-Till Rainfed 43.3 15.4 4.3.10 Genotypic, phenotypic and environment coefficient of variation for WUE under different tillage and moisture regimes The genotypic coefficient of variation for WUE was low in all the four environments (Table 4.8). Phenotypic coefficient of variation was moderate in all the environments whereas environmental coefficient of variation was low under irrigated conditions and moderate under rainfed conditions irrespective of the tillage regime (Table 4.8). Table 4.8: Coefficient of variation for WUE in different tillage and moisture regimes Parameter No-Till No-Till Till Till Irrigated Rainfed Irrigated Rainfed Genotypic coefficient of variation 5.8 7.4 9.4 8.8 Phenotypic coefficient of variation 13.9 17.2 13.9 14.2 Environmental coefficient of variation 9.0 14.1 8.3 11.6 4.4 Discussion The chickpea seed yields in the present study are similar to reports by Dalal et al. (1997) in their long-term experiment and Anwar et al. (2003) in their December sowing with full irrigation from flowering to podding. Supplementary irrigation applied twice at flowering and podding in this experiment increased seed yield with similar data reported by Silim and Saxena (1993) and Brown et al. (1989). Chickpea water use was higher in the no-till than the tillage system. This may be due to higher moisture availability under the no-till system since water use in chickpea depends on the levels of soil water available as observed by Singh and Bhushan (1980) in an experiment conducted 50 in Dehra Dun in northern India. The higher soil water levels under a no till system may be a result of increased soil moisture conservation and storage (Verhulst et al., 2010, Marley and Littler, 1989, Felton et al., 1995), reduced soil temperatures and evapotranspiration due to the presence of crop residues which lower wind speeds at the soil surface (Hatfield et al., 2001, Jones et al., 1994). It may also be due to increased infiltration levels due to the presence of macropores formed by roots of the previous crop and earthworms in the soil profile. Water use in this experiment ranged between 296 mm to 355 mm on average, which was within the reported range by Anwar et al. (1999). However, Singh and Bhushan (1980) reported a range of 109 mm to 208 mm for rainfed experiments which was lower than in the present study. This may be due to different soil types, climate and soil moisture availability levels. There was no difference in the water use of the genotypes tested. Similar data have been reported by Brown et al. (1989). The plants used water from the top 30 cm during the vegetative phase and later accessed soil moisture deeper in the horizon. This was also reported by Brown et al. (1983). Chickpea WUE was higher under no-till conditions than in the till system with similar findings reported by Herridge et al. (1995). WUE was higher under irrigated conditions than rainfed conditions. This was contrary to reports by Gan et al. (2010) who found that WUE was higher in rainfed conditions compared to irrigated conditions, whereas Anwar et al. (2003) did not find any significant differences between similar conditions. WUE was higher in 2014 than in 2015 which may have resulted from the weather conditions, especially rainfall distribution with higher rainfall in 2014 than 2015. More even rainfall distribution leads to better utilisation of soil moisture and consequently, higher grain yield. The diurnal range was lower in 2014 than in 2015 and plants also had deeper roots in 2014 than 2015 enabling them access to stored soil moisture from deep down the horizon. Such seasonal variation in WUE was also reported by Brown et al. (1989). Water use efficiency among individual genotypes ranged from 3.2 to 6.4 kg ha-1 mm-1 under different tillage and moisture treatments based on a two year average. These values were in the range of findings by Gan et al. (2010) who reported WUEs between 5.3 to 6.7 kg ha-1 mm-1 in Saskatchewan. Dalal et al. (1997) reported a mean WUE value of 5.9 kg ha-1 mm-1 at Warra in Queensland, while Herridge et al. (1995) reported a mean of 5.8 kg ha-1 mm-1 in Glenhoma, New South Wales, whereas Beech and Leach (1988) reported 4.9 kg ha-1 mm-1 in Dalby, south eastern Queensland. The minor differences between the values in the present study and the ones reported by the other authors may be partly attributed to how water use was measured. One of the challenges in the present study was that the neutron probe access tubes were not 51 inserted in all the plots but only in the control plots. Water use for the other genotypes apart from the control genotypes was calculated using the average water use of the control genotypes in each environment type. This may have underestimated or overestimated water use efficiency of the other tested genotypes. Genotypic variation in the tested genotypes was low; hence the need to diversify the genetic base of the materials in the present study through the introduction of new germplasm or hybridisation. Selection for genotypes with high WUE can be achieved under till systems in both rainfed and irrigated environments. This is because the tested genotypes showed high heritability and genetic advance scores under both tilled rainfed and irrigated systems. The high heritability recorded for plants grown under the till irrigated system may be associated with the breeding environment where these genotypes have been developed (Trethowan et al., 2012). 4.5 Conclusions Chickpea genotypes in the present study did not show variation for water use but they varied significantly in their WUE. This variation can be exploited by choosing suitable parents contrasting for WUE to start a hybridisation program. Water use efficiency can be improved by adapting the no till system with supplementary irrigation if water is not limiting. However, if supplementary irrigation is not feasible, there are still yield benefits of planting chickpea under a no till system. More research is needed to identify more sources of genetic variation in chickpea which will enable breeding programs to develop new varieties with high water use efficiency. 52 CHAPTER 5: THE BASIS OF CHICKPEA YIELD FORMATION UNDER WATER LIMITED FIELD CONDITIONS. 5.1 Introduction According to the Food and Agriculture Organization (FAO, 1983), drought is defined as ‘the percentage of years when crops fail from lack of moisture’. Agricultural drought is considered as a period where there is a decline in soil moisture which will eventually lead to crop failure (Mishra and Singh, 2010). The balance between plant water demand and supply is critical and water stress occurs if the demand outstrips the supply, especially in the top soil layer where there is a higher concentration of nutrients, soil microorganisms and root activity (Kulik, 1962). Water stress causes water limitation in the soil, hence limiting the amount of water available for crop growth and development. Water stress can occur at any point in the growing season with varied effects on the crop. Intermittent drought occurs at any time during the growing season due to rainfall breaks, whereas terminal drought occurs towards the end of the growing season due to a steady decline in soil moisture (Canci and Toker, 2009). Chickpea is mainly grown on residual soil moisture in Australia and, as such, suffers from terminal drought (Sedgley et al., 1990, Turner, 2003, Leport et al., 1998, Krishnamurthy et al., 2010) with up to 50% yield losses (Varshney et al., 2013b). However, these yield losses can range from 30% to 100% depending on the environment, type of drought and genotype (Leport et al., 1999). Drought tolerance research is imperative in identifying genotypes which can perform well in water limited environments. Sojka et al. (1981) defined drought tolerance as the ability of a plant to minimise yield losses under water limited conditions. Screening genotypes for yield potential and assessing their performance under differing moisture regimes is a key starting point in drought tolerance research (Ahmad et al., 2003). Various screening methodologies have been proposed which include selection of genotypes under water stressed conditions (Ceccarelli and Grando, 1991), selection under well-watered conditions (Betran et al., 2003, Richards, 1996) and selection under both well-watered and water stressed conditions (Fischer and Maurer, 1978, Clarke et al., 1992). Plant breeders are mainly guided by yield while selecting for drought tolerance, hence proper screening and selection is necessary (Ganjeali et al., 2011). Different selection indices, which compare yield loss in stress conditions to normal conditions, have been established to aid in 53 the selection for drought tolerance (Mitra, 2001, Farshadfar et al., 2013). These indices have been used to identify drought tolerant genotypes in chickpea (Ganjeali et al., 2011), wheat (Talebi et al., 2009, El-Rawy and Hassan, 2014), barley (Nazari and Pakniyat, 2010), sunflower (Gholinezhad et al., 2014, Darvishzadeh et al., 2011), and oats (Akcura and Ceri, 2011). The indices provided a weighted method of identifying drought tolerant genotypes without entirely relying on yield which can give erroneous result. Selecting genotypes that have high yield under water stressed conditions may also translate to high yield potential in well-watered conditions (Blum, 1988). However, Rosielle and Hamblin (1981) posited that selection of these genotypes under stress environments may not necessarily lead to high yields under well-watered conditions. A better approach would be to understand the physiological basis of yield formation under water limited environments and use these traits to select for drought tolerance. Selection of physiological traits, which are drought adaptive, coupled with high yield, potentially results in a plant with high yield and physiological shock absorbers against drought (Blum, 1983). The identification of traits of interest that can be used as an indirect selection criterion in a breeding or introgression program is referred to as physiological breeding (Jackson et al., 1996). Traits can be identified by using either the black box approach or the ideotype approach. In the black box approach, genotypes are evaluated under stress conditions and trait association with economic performance is measured, whereas under the ideotype approach, the desired traits in an ideal genotype in a given target environment are predicted (Fischer, 1981). In both methods, the target environment should be as close as possible to the treatments being administered in all the plots to minimise the occurrence of confounding factors that can affect trait expression (Reynolds et al., 2001). By using physiological breeding strategies (Reynolds et al., 2012), coupled with selecting for drought tolerance using drought indices, one can identify drought tolerant genotypes that can perform well in water stressed and well-watered conditions. The research question is whether certain traits can be selected to confer drought tolerance in chickpea, and whether drought tolerance indices can be used to identify these marker traits in the field? The aims of this chapter are to: i) identify drought tolerant and drought susceptible genotypes that can either be grown directly by farmers or used as parents in a breeding program, ii) identify the phenological, morphological and physiological basis of drought tolerance, iii) identify the most suitable selection indices for drought tolerance in chickpea, and iv) identify 54 phenological, morphological and physiological marker traits to aid selection of drought tolerant genotypes during the active growing season. 5.2 Materials and methods A total of 36 entries were grown over two years (2014 and 2015) at the IA Watson Plant Breeding Institute in Narrabri, as described in the general Materials and Methods in Chapter 3. For the purposes of this chapter, only the well-watered and water stressed treatments are considered because yield differences between no-till and till were minimal, and there was no tillage by genotype interaction effect. Data on phenology, physiological and morphological traits were recorded from the vegetative stage through to maturity during the growing season. Grain was harvested from a 4 m by 2 m plot for each entry and data recorded. Data recorded included percent early ground cover. This was done by taking pictures using a NikonTM camera of the plot to cover four rows and then it was analysed using the CSIRO Canopy Cover Software to give percent ground cover. Days to first flower was recorded as the day the first open flower was sighted in a plot. Days to 50% flowering was recorded as day when 50% of the plants in a plot have at least one open flower. Days to last flower was recorded as the day when the last open flower was sighted in the plot. Flower duration was calculated as days to last flower minus the days to first flower. Normalised Difference Vegetation Index (NDVI) measurements were done using a GreenseekerTM machine at early podding and late podding. Plant height was measured using a ruler from the base of the stem at ground level to the tallest growing tip at late flowering and late podding. Chlorophyll content was measured using a SPAD meter at mid and late podding stages. Number of leaflets per leaf was counted by getting an average of 10 fully grown leaves per plot sampled at the fifth leaf from the top at flowering. Single leaf area was obtained by measuring with a ruler average leaf area from 5 leaves sampled from 5 plants, 1 leaf per plant at flowering. Single leaflet size and length were measured using a ruler as an average of 10 fully grown leaves per plot sampled at the fifth leaf from the top at flowering. Number of pods per plant was obtained by counting and obtaining the mean of 5 representative plants per plot at maturity. Pod biomass was obtained by taking a sample of 5 plants in a plot, drying them and weighing their mass and then getting the average mass of the 5 plants. Pod harvest index was calculated as seed yield from pod biomass samples divided by pod biomass whereas shoot harvest index was calculated by grain yield divided by biological yield. One thousand weight was obtained by sampling 100 seeds at 10% moisture content and weighing them and multiplying by 10. 55 Various drought susceptibility indices (Table 5.1) were calculated from the yield data collected. They included mean relative performance (Rosielle and Hamblin, 1981, Reddy et al., 2009) where a higher value denotes tolerance, and relative efficiency index (Singh et al., 2011) which selects genotypes with high yield potential and are drought tolerant with high values being desirable. High values according to the stress tolerance index (Fernandez, 1992), drought resistance index (Lan, 1998), yield index (Gavuzzi et al., 1997) and yield stability index (Bouslama and Schapaugh, 1984) are correlated with drought tolerance. Linear regression was run for the drought indices and yield using the formula Where Yi is the score for the dependent variable for the ith term, a + b Xi are a linear function relating X (of the ith term) to Y, and e i is the error term. Heritability estimates were estimated using the formula in equation 5.1 (Knapp and Bridges, 1987). H = (σ2g) / (σ2g + (σ2ge/e) + (σ2e/re)) (5.1) Where H is the broadsense heritability, σ2g is the genotypic variance, σ2ge is the genotype by environment interaction variance, σ2e is the error variance, e is the number of environments (years) and r is the number of replications. Genetic advance (GA) was calculated as shown in the equation 5.2 (Singh and Chaudhary, 1979) below and then converted to a percentage of the mean. GA = ((K*√σ2p* H2)/ X̅) * 100 (5.2) Where √σ2p is the phenotypic standard deviation, H2 is the heritability and K is the selection differential at 5% selection intensity (2.06). Yield data was analysed using Genstat® edition 18 by subjecting it to generalised linear mixed models (GLM). Tillage regime, water regime and genotypes were fitted in the fixed model whereas the range and row nested into years were fitted in the random model. Mean yield for well-watered and water stressed conditions were tabulated and used to calculate the drought indices. Physiological and morphological traits explaining most of the variation under water limited conditions were identified by regressing the traits against grain yield in multiple linear regressions. All traits that did not significantly explain variation in yield in each run were 56 eliminated until all the traits that remained were significant at the 95% confidence interval. Selection of drought indices was done using principal component analysis and plotting a principal component scatter plot to observe their relationships with each other and the genotypes. Table 5.1: Drought tolerance indices for evaluating chickpea yield under water-limiting conditions Index Abbreviation Equation Equation no. Mean Relative Performance MRP (Ysi/Ys) + (Ypi/Yp) (5.3) Relative Efficiency Index REI (Ysi/Ys) * (Ypi/Yp) (5.4) Stress Tolerance Index STI (Ysi * Ypi) / (Yp)2 (5.5) Drought Resistance Index DRI (Ysi *(Ysi/Ypi)) / (Ys) (5.6) Yield Index YI Ysi / Ys (5.7) Yield Stability Index YSI Ys / Yp (5.8) Where Ysi is the yield under stress for the ith genotype, Ypi is the yield under well-watered conditions for the ith genotype, Ys is the mean grain yield under stress conditions and Yp is the mean grain yield under well-watered conditions. 5.3 Results 5.3.1 Phenological, morphological and physiological traits for yield formation under water stressed conditions Twenty-one traits accounted for 91% of the total variation in yield from the multiple linear regression (Table 5.2). The traits included phenological, morphological, physiological and yield components with confidence levels ranging from p<0.05 to p<0.001. Important phenological traits include days to first flower, days to 50% flowering and days to last flower. Flowering is important, especially in areas where there is water limitation towards the end of the growing season. Early flowering ensures there is adequate soil moisture at the reproductive phase in contrast to late flowering where there is high risk of soil water deficit and a loss in yield potential. Important morphological traits included leaf characteristics and plant height. Leaf area plays a key role in water loss through the transpiration stream – large leaf surface areas lose more water compared with small surface areas. Important physiological traits include NDVI, chlorophyll content and early ground cover. A high NDVI during the reproductive phase was associated with high yield, however a high NDVI towards the end of the growing season was associated with low yield. Similar to NDVI, a high chlorophyll content at mid podding was associated with high yield and low yield towards the end of the growing season. 57 Early season ground cover resulted in high yields at the end of the season. Yield component traits that explained much of the variation in grain yield included the number of pods per plant, pod and shoot biomass, pod and shoot harvest index and 1000 seed weight. Moderate biomass for pod and shoot, as well as moderate seed weight, resulted in high yields at the end of the growing season. A high harvest index is desirable since it resulted in high yields as well. All the traits measured had high heritability (greater than 60%) apart from NDVI at early podding which had a low heritability of 43% (Table 5.2). Thousand seed weight had the highest heritability of 99% closely followed by morphological traits (leaf characteristics) and phenological traits (days to first flower, days to 50% flowering and flower duration, except days to last flower which was lower than the rest). Physiological traits and yield component traits had notably lower heritability estimates compared with phenological and morphological traits. Early ground cover, flowering duration, NDVI at late podding, leaf area and leaflet length, number of pods per plant, pod biomass, shoot biomass, shoot harvest index and one thousand seed weight had high heritability estimates and genetic advance. The lowest genetic advance was recorded for the pod harvest index and NDVI at early podding stage. Table 5.2: Traits explaining variation in chickpea yield under water stressed conditions, correlations with grain yield, heritability and genetic advance Wald Trait statistic Correlation Heritability GA (%) Early ground cover (%) 14.2** 0.10 79.6 50.5 Days to first flower 31.66*** -0.18 97.0 14.2 Days to 50% flowering 18.63*** -0.23 95.3 12.4 Days to last flower 24.69*** -0.36* 78.0 2.6 Flowering duration (days) 24.62*** 0.12 95.2 29.7 NDVI at early podding 58.92*** 0.55* 43.0 4.7 NDVI at late podding 8.32* -0.52* 83.7 31.0 Plant height at late flowering 20.72*** 0.10 88.9 14.5 Plant height at late podding 29.84*** 0.04 92.5 14.9 Chlorophyll content at mid podding (SU) 5.38* 0.02 71.7 7.5 Chlorophyll content at late podding (SU) 9.61** -0.35* 73.3 16.6 Number of leaflets per leaf 13.42** -0.10 96.3 11.7 Single leaf area (cm2) 25.64*** -0.04 98.3 89.7 58 Single leaflet area (cm2) 35.79*** -0.03 98.4 91.0 Leaflet length (cm2) 28.16*** 0.02 97.7 37.4 Number of pods per plant 11.43** -0.18 74.7 50.4 Pod biomass per plant (g) 19.11*** -0.31 85.7 82.7 Pod harvest index 9.6** 0.10 56.2 3.5 Shoot biomass 25.85*** -0.41* 84.7 67.2 Shoot harvest index 47.88*** 0.24 84.5 22.9 One thousand seed weight (g) 9.05** -0.39* 99.1 53.1 Where SU is SPAD Units, * = P<0.05, ** = P<0.01 and *** = P<0.001 5.3.2 Grain yield and drought indices Grain yield ranged from 1222 kg ha-1 to 2074 kg ha-1 under well-watered conditions and 1170 kg ha-1 to 1850 kg ha-1 under water stressed conditions (Table 5.3). On average, the wellwatered moisture regime resulted in a higher yield (1722 kg ha-1) than the water stressed moisture regime (1478 kg ha-1) with drought causing a 14% reduction in grain yield. The highest yielding genotype under well-watered conditions was PBA Slasher followed by Sonali, whereas under water limited conditions, Sonali was the highest yielding genotype. By ranking the genotypes based on their performance in well-watered and water stressed conditions, Sonali, PBA Slasher, ICCV 96853 and Jimbour were classified as stable, whereas Amethyst dropped from a ranking of 7 to 30, and Genesis 079 dropped from a ranking of 14 to 31 under well-watered and water stressed conditions, respectively (Table 5.3). This demonstrates that Amethyst and Genesis 079 have high yield potential but are vulnerable to water stressed conditions, hence the high loss in grain yield. Based on the grain yield ranking and drought indices, PBA Slasher, Sonali, ICCV 96853 and Jimbour were identified as drought tolerant, whereas Amethyst and Genesis 079 were drought susceptible. All the indices ranked Sonali as the most tolerant genotype except for the yield stability index which ranked it seventh. 59 Table 5.3: Grain yield and drought tolerance indices for chickpea genotypes grown under well-watered and water stressed conditions Grain Grain yield yield (kgha-1) (kgha-1) (WW) (WS) PBA SLASHER 2074 SONALI Rank Rank under under WW WS 1657 1 2041 1850 ICCV 96853 2020 JIMBOUR Genotype MRP REI STI DI YI YSI 3 2.33 1.35 1.16 0.77 1.12 0.75 2 1 2.44 1.48 1.27 0.97 1.25 0.88 1714 3 2 2.33 1.36 1.17 0.84 1.16 0.82 1873 1643 4 4 2.20 1.21 1.04 0.84 1.11 0.91 PBA HATTRICK 1855 1630 5 6 2.18 1.19 1.02 0.83 1.10 0.80 Mix 5 (Howzat/Jimbour) 1850 1605 6 8 2.16 1.17 1.00 0.81 1.09 0.77 AMETHYST 1818 1372 7 30 1.98 0.98 0.84 0.60 0.93 0.81 TYSON 1806 1562 8 11 2.11 1.11 0.95 0.78 1.06 0.88 Mix 6 (Howzat/ 98813) 1800 1632 9 5 2.15 1.15 0.99 0.86 1.10 0.85 Mix 4 (Yorker/Howzat) 1799 1441 10 22 2.02 1.02 0.87 0.67 0.97 0.93 JIMBOUR#1 1791 1503 11 14 2.06 1.06 0.91 0.73 1.02 1.07 Mix 2 (Howzat/Flipper) 1771 1455 12 18 2.01 1.01 0.87 0.69 0.98 0.85 Mix 3 (Flipper/Jimbour) 1771 1540 13 12 2.07 1.07 0.92 0.78 1.04 0.92 GENESIS 079 1756 1352 14 31 1.93 0.93 0.80 0.60 0.91 0.83 LYONS 1751 1473 15 16 2.01 1.01 0.87 0.72 1.00 0.90 DOOLIN 1749 1441 16 21 1.99 0.99 0.85 0.69 0.97 0.92 HOWZAT 1743 1614 17 7 2.10 1.10 0.95 0.87 1.09 0.88 ICCV 98801 1738 1598 18 9 2.09 1.09 0.94 0.85 1.08 0.84 Mix 1 (Yorker/Jimbour) 1735 1445 19 20 1.98 0.98 0.85 0.70 0.98 0.89 LYLE 1727 1440 20 23 1.98 0.98 0.84 0.70 0.97 0.83 PBA STRIKER 1727 1589 21 10 2.08 1.08 0.92 0.85 1.07 0.84 SIM 1723 1452 22 19 1.98 0.98 0.84 0.71 0.98 0.83 THOMAS 1713 1434 23 24 1.96 0.96 0.83 0.70 0.97 0.82 KYABRA 1695 1514 24 13 2.01 1.01 0.87 0.79 1.02 0.87 GENESIS 090 1688 1374 25 29 1.91 0.91 0.78 0.65 0.93 0.80 YORKER 1666 1392 26 26 1.91 0.91 0.78 0.68 0.94 0.87 HOWARD 1663 1415 27 25 1.92 0.92 0.79 0.70 0.96 0.91 AUSTIN 1661 1459 28 17 1.95 0.95 0.82 0.74 0.99 0.88 FLIPPER 1659 1322 29 33 1.86 0.86 0.74 0.61 0.89 0.80 FLIP 079C 1642 1489 30 15 1.96 0.96 0.82 0.78 1.01 0.92 SAL 1628 1348 31 32 1.86 0.86 0.74 0.65 0.91 0.83 ICCV 98816 1544 1385 32 28 1.83 0.84 0.72 0.72 0.94 0.84 ICCV 98818 1510 1387 33 27 1.82 0.82 0.71 0.74 0.94 0.91 ICCV 98813 1469 1216 34 35 1.68 0.70 0.60 0.58 0.82 0.84 60 GENESIS KALKEE 1328 1170 35 36 1.56 0.61 0.52 0.60 0.79 0.87 ICCV 05308 1222 1309 36 34 1.59 0.63 0.54 0.81 0.89 0.84 Where WW is well watered, WS is water stressed, MRP is mean relative performance, REI is relative efficiency index, STI is stress tolerance index, DI is drought resistance index and YI is yield index. 5.3.3 Grain yield relationships under well-watered and water stressed conditions There was a moderately strong positive relationship between grain yield under well-watered and water stressed conditions (Figure 5.1). This shows that genotypes with the highest yield potential under well-watered conditions generally yielded well under water stressed conditions as well although there were some exceptions. This sort of plasticity is important in plant -1 I r r ig a te d Y ie ld (k g h a ) breeding programs. 2000 1500 R 2 = 0 .6 5 Y = 0 .9 7 2 7 * X + 2 8 4 .4 1000 1000 1500 R a in fe d Y ie ld (k g h a 2000 -1 ) Figure 5.1: Relationship between irrigated (well-watered) and rainfed (water stress) yield for the different chickpea genotypes analysed for drought tolerance 5.3.4 Correlation analysis for grain yield and drought indices All the slopes had a significant (P<0.05) deviation from zero except for yield stability index which was non-significant (P>0.05). The intercepts and slopes were also different. The mean relative performance index was highly and positively correlated with grain yield for both wellwatered and water stressed conditions with a coefficient of determination (R2) of 0.91 and 0.90, respectively (Figure 5.2a). This implies selecting for genotypes with a high relative performance index will lead to high yield potential. Both relative efficiency index and stress tolerance index have similar coefficient of determinations for well-watered and water stressed 61 conditions. Under well-watered conditions the R2 was slightly lower at 0.89 than under water stressed conditions (R2 = 0.91) (Figure 5.2b and 5.2c). The drought resistance index had a weak and positive relationship with grain yield under well-watered conditions, however it was not significant (Figure 5.2d). Still, this index had a high and positive relationship with grain yield under water stressed conditions suggesting its suitability for selection under drought stress conditions. The yield index had a moderate positive correlation with grain yield, and a strong positive correlation with grain yield under well-watered and water stressed conditions respectively (Figure 5.2e). However, the yield stability index was not associated with grain yield in either well-watered or water stressed conditions (Figure 5.2f). R 2 a = 0 .9 1 c b Y = 873.4*X - 24.43 2000 R 2 = 0 .9 0 G r a in y ie ld (k g h a -1 ) Y = 722.6*X + 33.18 1500 2 R = 0 .8 9 2 R = 0 .8 9 Y = 867.5*X + 849.1 2 R Y = 1009*X + 849.6 R 2 = 0 .9 1 = 0 .9 1 Y = 848.1*X + 745 Y = 728.8*X + 744.8 1000 e d f 2000 2 1500 R = 0 .0 0 9 3 2 Y = -285.1*X + 1968 R 2 = 0 .1 7 R = 0 .6 4 Y = 767.9*X + 1155 Y = 1429*X + 293.9 2 R = 0 .0 0 0 0 0 8 7 2 R = 0 .9 9 2 R = 0 .7 6 Y = -7.232*X + 1485 Y = 1478*X + 1.106 Y = 1330*X + 495.1 1000 0 1 2 0 1 2 0 1 2 D r o u g h t in d e x Figure 5.2: Linear regression of chickpea grain yield against six drought indices shown for well-watered (blue) and water-stressed (red) growing conditions. Blue data points represent well-watered conditions and red data points represent water stressed conditions. Each plot represents grain yield on the Y-axis and a specific drought index on X-axis where a) is mean relative performance, b) is relative efficiency index, c) stress tolerance index, d) is drought resistance index, e) is yield index and f) is yield stability index. 62 5.3.5 Selection of the best drought tolerance index for chickpea The first principal component explained 92.72% of the total variation in drought tolerance, whereas the second principal component explained 4.47% of the total variation (Figure 5.3). Genotypes 7, 9 and 14 (PBA Slasher, Sonali and ICCV 96853, respectively) clustered near each other and are considered drought tolerant, whereas the drought susceptible entries 1 and 12 (Amethyst and Genesis 079, respectively) clustered together when analysed by the different drought indices (Figure 5.3). The drought response index, stress tolerance index, mean relative performance and relative efficiency index were the most discriminating for identifying drought tolerant genotypes, whereas the yield index and yield stability index were not able to efficiently identify drought tolerant genotypes. The stress tolerance index, mean relative performance and relative efficiency indices were positively correlated to each other. These indices were also positively correlated with the other indices except for the yield stability index. 63 Figure 5.3: Principal component scatter plot for chickpea genotypes and drought indices. MRP, mean relative performance; REI, relative efficiency index; STI, stress tolerance index; DI, drought resistance index; YI, yield index. The green oval shape groups the drought tolerant genotypes and the orange oval shape groups drought susceptible genotypes. 5.3.6 Effect of water deficit on important traits associated with chickpea grain yield under water limited conditions In general, water deficit caused the reduction of trait means except for a few instances where the means increased. Early ground cover had a mean of 20.4% which was reduced by 7% under water stressed conditions compared with well-watered conditions (Table 5.4). Most phenological traits had minimal change except in flowering duration which decreased by 13.2%, thus; genotypes under well-watered conditions flowered almost five days later 64 compared with water stressed conditions. There was a 1.9% reduction in NDVI during early podding compared with a 42.2% reduction in NDVI at late podding in water stressed compared with well-watered conditions. This is probably indicative of the difference in soil moisture from early podding to the late podding stages. Plant height was shorter under water stressed than well-watered conditions with a 5.7% and 9.1% reduction at the mid and late podding stages, respectively. On the contrary, chlorophyll content at the mid and late podding stages increased by 4.1% and 1.9%, respectively, under water stressed conditions. Number of leaflets per leaf also increased from an average of 13.7 under well-watered conditions, to 14 under water stressed conditions, denoting a 1.9% increase. Other leaf traits including single leaf area, single leaflet area and leaflet length decreased by 10.2%, 12.3% and 4.6%, respectively, under water stressed conditions. Yield components were more affected by water stress compared with the other traits analysed. The numbers of pods per plant were reduced from 38.9 to 31.3 when plants were exposed to water stressed conditions. Pod biomass and shoot biomass were reduced by 28.4% and 25.1%, respectively, under water deficit conditions. The pod harvest index was similar between the treatments (0.79 to 0.80) whereas and the shoot harvest index reduced by 2.2% by the water stress conditions. One thousand seed weight increased from 212 g to 217 g under water stressed conditions denoting a 2.2% increase (Table 5.4). Table 5.4: Trait means and per cent change due to water deficit in chickpea genotypes Trait Trait mean Trait mean Change (WW) (WS) (%) Early ground cover (%) 20.4 19.1 -7.0 Days to first flower 87.4 86.6 -1.0 Days to 50% flowering 98.2 97.2 -1.1 Days to last flower 127.6 122.0 -4.5 Flowering duration 40.2 35.5 -13.2 NDVI at early podding 0.7 0.7 -1.9 NDVI at late podding 0.5 0.3 -42.2 Plant height at late flowering 62.9 59.5 -5.7 Plant height at late podding 70.0 64.1 -9.1 Chlorophyll content at mid podding (SU) 67.7 70.6 4.1 Chlorophyll content at late podding (SU) 56.9 58.0 1.9 Number of leaflets per leaf 13.7 14.0 1.9 65 Single leaf area (cm2) 7.7 7.0 -10.2 Single leaflet area (cm2) 0.6 0.5 -12.3 Leaflet length 1.3 1.2 -4.6 No of pods per plant 38.9 31.3 -24.4 Pod biomass per plant (g) 13.9 10.9 -28.4 Pod harvest index 0.8 0.8 1.1 Shoot biomass (cm2) 28.3 22.6 -25.1 Shoot harvest index 0.4 0.4 -2.2 One thousand seed weight 212.0 216.8 2.2 Where SU is SPAD Units and NDVI is normalised difference vegetation index. 5.3.7 Associations between trait relationships and chickpea drought indices The phenological, morphological, physiological and yield component traits were associated with drought indices either positively or negatively but not all associations were significant (Table 5.5). Days to 50% flowering and days to last flower had a negative and significant correlation with drought resistance index but did not have any significant relationship with the other drought indices. The NDVI at early podding was significantly and positively associated with all the indices, except for drought resistance index, where the association was not significant, and yield stability index where the association was negative (Table 5.5). The NDVI at late podding had a negative and significant relationship with all the indices, except for yield stability index, where the association was not significant at the 95% confidence interval. The chlorophyll content at late podding was significantly and negatively correlated with all the indices except drought response index and yield stability index. However, chlorophyll content at mid podding was not significantly associated with any of the drought indices. Leaflet length had a positive and significant relationship with the drought response index only. Pod biomass per plant had a significant and negative correlation with mean relative performance, relative efficiency index, and stress tolerance index, and a non-significant relationship with yield index and yield stability index. Similar relationships were observed between these indices and shoot biomass also except the yield index was significantly and negatively related. Thousand seed mass was significantly and negatively correlated with all the drought indices except for drought resistance index and yield stability index. In general, mean relative performance, relative efficiency index and stress tolerance index were correlated with the indices in a similar fashion closely followed by the yield index. These indices were mainly correlated with physiological traits and yield components. The drought resistance index exhibited a different trend from the 66 other indices but was significantly correlated with phenology (days to 50% flowering and days to last flower), morphological traits (leaflet length), physiological traits (NDVI at late podding). The yield stability index was not significantly correlated with any trait. Table 5.5: Correlation between drought indices and important traits in chickpea grown under water deficit conditions Trait MRP REI STI YI DI YSI Early ground cover 0.14 0.14 0.14 0.10 0.01 0.04 Days to first flower -0.09 -0.11 -0.11 -0.18 -0.32 0.08 Days to 50% flowering -0.14 -0.16 -0.16 -0.23 -0.35 0.06 Days to last flower -0.27 -0.29 -0.29 -0.36 -0.44 0.14 Flower duration 0.03 0.05 0.05 0.12 0.26 -0.05 NDVI at early podding 0.62 0.61 0.61 0.55 0.31 -0.17 NDVI at late podding -0.56 -0.56 -0.56 -0.52 -0.36 -0.08 Plant height at late flowering 0.15 0.12 0.12 0.10 -0.02 0.10 Plant height at late podding 0.06 0.04 0.04 0.04 -0.03 -0.02 Chlorophyll content at mid podding (SU) -0.03 -0.03 -0.03 0.02 0.08 0.10 Chlorophyll content at late podding (SU) -0.42 -0.40 -0.40 -0.35 -0.17 0.06 Number of leaflets per leaf -0.11 -0.10 -0.10 -0.10 -0.07 -0.19 Single leaf area (cm2) -0.24 -0.20 -0.20 -0.04 0.31 -0.09 Single leaflet area (cm2) -0.23 -0.19 -0.19 -0.03 0.32 -0.07 Leaflet length -0.17 -0.13 -0.13 0.02 0.34 -0.08 Number of pods per plant -0.22 -0.22 -0.22 -0.18 -0.03 -0.01 Pod biomass per plant (g) -0.48 -0.45 -0.45 -0.31 0.05 -0.07 Pod harvest index 0.29 0.26 0.26 0.10 -0.23 0.01 Shoot biomass (cm2) -0.56 -0.54 -0.54 -0.41 -0.05 -0.02 Shoot harvest index 0.19 0.21 0.21 0.24 0.30 -0.07 1000 seed mass -0.55 -0.51 -0.51 -0.39 -0.06 -0.09 MRP, mean relative performance; REI, relative efficiency index; STI, stress tolerance index; DI, drought resistance index; YI, yield index; YSI, yield stability index; SU, SPAD units. Figures in bold indicate significance at a 95% confidence interval (p<0.05). 5.4 Discussion Plant breeders and growers require genotypes that are high yielding in non-stress conditions and have minimal yield losses under stress conditions (Ud-Din et al., 1992). It is therefore 67 imperative to consider rankings coupled with selective indices because a single drought tolerance selection criteria may be misleading (Khalili et al., 2012). There was a significant correlation (R2 = 0.65) between non-stress yield and water-stress yield suggesting that some genotypes that yield highly under non-stress conditions also yield highly under stress conditions. This means that direct selection of chickpea genotypes under non-stress conditions may be a predictor of good performance under stress conditions. The best drought tolerance indices should have a high correlation with both non-stress yield and stress yield (Mitra, 2001, Blum, 1988). This helps in selecting genotypes which show plasticity such that in years with adequate rainfall, these genotypes take advantage of adequate moisture and give higher yields, and in years with low rainfall, they still produce and do not experience total crop failure. The relative efficiency index, stress tolerance index, drought resistance index, yield index had stronger correlations with yield in stressed plants compared with non-stress yield. These findings concur with those reported by (Sahar et al., 2016). The relative efficiency index, stress tolerance index, drought resistance index and yield index had a significant and positive relationship with yield and this is in agreement with findings by Kumar et al. (2014), Sahar et al. (2016) and Singh et al. (2011) in studies conducted in rice, bread wheat and sorghum, respectively. Stress tolerance index had a positive and significant correlation with non-stress yield with a coefficient of determination of 0.91. Similar data were reported by Nazari and Pakniyat (2010) in barley genotypes whereas Talebi et al. (2009) reported an R2 of 0.79 in durum wheat which was slightly lower. Stress tolerance index was a better predictor of drought tolerance than mean relative performance (Talebi et al., 2009, Nazari and Pakniyat, 2010) which is in agreement with data reported in the present study. This is because stress tolerance index was highly correlated with yield and it identified genotypes with high yield potential and high drought tolerance. Stress tolerance index and relative efficiency index were identical based on the biplot analysis indicating similar genotype rankings with respect to drought tolerance. Based on the biplot constructed from the principal component analysis (PC1 and PC2), Mean relative performance, relative efficiency index and stress tolerance index were closely related and were the best predictors to identify drought tolerant genotypes in chickpea with high yield potential. Since grain yield is highly affected by the interaction between genotype and environment, it is more practical to identify traits that are associated with yield under water stressed environments and use them as selection criteria (Ludlow and Muchow, 1990). Twenty one phenological, morphological and physiological traits were identified as important in explaining yield 68 variation under water stressed conditions (Table 5.3). Ramamoorthy et al. (2016) identified days to 50% flowering, shoot biomass at maturity, harvest index and number of pods as important traits under water stressed conditions. These are among the 21 traits identified in the present study. Early ground cover was associated with high yielding genotypes at the end of the growing season. This may be attributed to the fact that there is high moisture loss from the ground as a result of evaporation, hence early ground cover reduces these losses (Siddique et al., 2001). Early flowering genotypes performed better than late maturing genotypes in the present study because genotypes that mature early are able to avoid terminal drought at the end of the growing season (Toker et al., 2007). Genotypes with high NDVI values at the early podding stage and high chlorophyll content at the mid podding stage had high yields and were drought tolerant. These data are similar to other reports (Maalouf et al., 2011). However, genotypes that had high NDVI values and chlorophyll content values towards the end of the growing season were low yielding under water limited conditions. Genotypes with small leaves had higher yields under water stressed environments probably due to reduced evaporative surface area thereby conserving water in the soil. This may also be attributed to the fact that smaller leaves contribute to an increased rate of partitioning to grains (Ramamoorthy et al., 2016). Low shoot biomass in the present study was associated with low yields under water stressed conditions. This was contrary to reports by Kashiwagi et al. (2015) who found higher shoot biomass led to high yields and better drought tolerance under water stressed conditions. This discrepancy may be due to the different genotypes used in the experiments and the differences in the environments. The shoot harvest index indicates the ability of a plant to partition assimilates and the reallocation of stored assimilates into grain yield (Turner et al., 2001). Shoot harvest index, number of pods per plant and days to first flower are important traits to consider under water limited conditions with similar findings reported by (Toker and Canci, 2005). Mean shoot harvest index was 0.39, which is slightly lower than the 0.42 reported by Siddique et al. (2001) in an experiment conducted at Mullewa in Western Australia. Genotypes with high shoot harvest index are normally high yielding under water stressed conditions (Krishnamurthy et al., 2013, Beebe et al., 2008). High heritability coupled with genetic advance is favourable because it implies additive gene action or cumulative contribution of alleles in the formation of a phenotype), whereby the effect 69 of environment on genotype is minimal. Selecting for such traits is attractive to the plant breeder because it means faster genetic gains can be made. Days to 50% flowering had high heritability which is similar to data reported by Ramamoorthy et al. (2016). The 1000 seed mass had the highest heritability in the present study which is similar to findings by Hamwieh and Imtiaz (2015) and Ramamoorthy et al. (2016) who reported heritability estimates of 8497% and 96% under water stressed conditions, respectively. Hay (1995) reported a high heritability estimate for harvest index which is similar to data in the present study. Water deficit reduced the expression of most traits except for chlorophyll content, number of leaflets per leaf, pod harvest index and 1000 seed mass. Days to first flower, days to 50% flowering and days to last flower were earlier than in the non-stress conditions with similar findings reported by (Ramamoorthy et al., 2016). There was a slight reduction in NDVI at the early podding stage, however, there was a 42% difference in NDVI at the late podding stage between water stressed and well-watered conditions. The lower NDVI at late podding may be attributed to leaf senescence due to water stressed conditions. Plants were shorter under water stressed conditions than in well-watered conditions most likely due to a reduction in cell expansion and enlargement due to low plant water status (Manivannan et al., 2007). There was a reduction in leaf area and number of pods per plant under water stressed conditions compared with the well-watered conditions and similar findings were reported by Randhawa et al. (2014). Water stressed conditions caused a 25% reduction in shoot biomass which may be attributed to reduced cell division as a result of impaired cyclin dependent kinase activity (Schuppler et al., 1998a). Since drought tolerance indices are based on yield, one has to wait until harvesting is completed to compute them and select the genotypes which show tolerance (El-Hendawy et al., 2017). However, by use of traits that are associated with these indices in the field during the active crop growth period, plant breeders can engage in early selection of promising genotypes which will eventually be drought tolerant. NDVI was positively and significantly correlated with mean relative performance, relative efficiency index, stress tolerance index and yield index at the early and late podding stages. A similar finding was reported by El-Hendawy et al. (2017) in an experiment using spring wheat lines where NDVI was correlated with yield index and stress tolerance index . Other stress tolerance indicator traits based on mean relative performance, relative efficiency index, stress tolerance index and yield index include chlorophyll content at late podding which can be measured during the active crop growth period. Days to 50% flowering, days to last flower and leaflet length can be used as a proxy 70 for drought resistance index but may not provide adequate information since drought resistance index is not as accurate in identifying drought tolerant chickpea genotypes with high yield potential. Yield components (pod biomass, shoot biomass and 1000 seed mass) are analysed at the same time as grain and can be used to confirm the crop performance with respect to drought tolerance at the end of the season. The identification of traits associated with drought tolerance in the field gives plant breeders an ability to engage in early selection of drought tolerant genotypes while still actively growing in the field. This also helps the breeder look at other market preferred traits as well as agronomic appearance of the genotype to help in decision making. 5.5 Conclusions Grain yield under water stressed and well-watered conditions was positively correlated. Hence, selection for high yield potential under similar environments can lead to high yields under water stressed conditions. Sonali, PBA Slasher and ICCV 96853 were identified as drought tolerant genotypes with high yield potential. These genotypes can be used as parents in a chickpea breeding program to improve drought tolerance of existing commercial cultivars, or grown directly by farmers since they are released varieties. Growing these genotypes can give high yields under well-watered conditions and have low yield penalty under water stressed conditions, providing more profitability and risk mitigation for the grower. Use of yield ranking scores coupled with drought indices is recommended for the identification of high potential genotypes with drought tolerance. In the present study, mean relative performance, relative efficiency index and stress tolerance index were identified as the best indices for identifying drought tolerant chickpea genotypes. High heritability coupled with genetic advance can be used to identify traits that are controlled by additive gene action. Some of these traits include flowering duration, early ground cover, NDVI at late podding, number of pods per plant, leaf area and leaflet length, shoot biomass, pod biomass, 1000 seed mass and shoot harvest index. Water stress reduced the expression of several characters with NDVI at late podding, number of pods per plant, pod and shoot biomass being the traits most affected. Several traits were identified as markers for drought tolerance during the active chickpea growing season. By using NDVI at the early podding and late podding stages, as well as chlorophyll content at late podding, one can identify genotypes with potentially high yield and high drought tolerance. 71 CHAPTER 6: EFFECT OF GENOTYPE BY ENVIRONMENT BY MANAGEMENT INTERACTIONS ON CHICKPEA PHENOTYPIC STABILITY 6.1 Introduction In Australia, the main chickpea growing regions are Northern New South Wales and Queensland which have a sub-tropical climate, and Western Australia which has a Mediterranean climate (Wells, 2013). It is also grown in smaller acreages in South Australia and Victoria. Narrabri, which is in northwest New South Wales, has a summer dominant rainfall with a median annual rainfall ranging from 600-800 mm (Dang et al., 2015). Chickpea in this area is grown during winter which is characterised by high temperatures towards the end of the growing season as well as low and variable in-season rainfall (Freebairn et al., 1991).This variance makes repeatability of yield results difficult and a challenge for plant breeders and growers alike who aim to sustain yields under various growing environments. In an effort to stabilise yield results, various management options are adopted including no-till practices and supplementary irrigation. No-till systems are beneficial because they improve soil structure (Page et al., 2013), increase soil aggregate stability (Chan and Mead, 1988, Li et al., 2007) and improve water storage (Radford et al., 1995, Felton et al., 1995) as a result of increased infiltration rates and reduced water evaporation. The increase in infiltration rate may be attributed to earthworm activity and plant roots from the previous crop which creates continuity of macropores. While certain adaptable genotypes can perform well in a diverse range of environments, some only perform well in specific environments. The mean performance of a genotype denotes its average performance whereas the stability measure indicates its variability across a number of environments (Yan et al., 2001). The lack of stability in yield is influenced by the genotype by environment interaction (Fox and Geiger). This interaction complicates the selection of genotypes in a breeding program by offsetting expected responses (Pande et al., 2013, Gauch and Zobel, 1997). It is further complicated by the fact that not all stable genotypes are high yielding across a wide range of environments; thus, they may be highly stable but have low yield potential implying stability alone is not necessarily a good thing. To dissect the Genotype x Environment Interaction (GxE), it is imperative to perform multienvironments trials (MET) which entail growing genotypes across a wide range of environments (Annicchiarico, 2002). Multi-environment trials data analysis allows one to 72 decipher the relationships between environments and explores the possibility of grouping these environments into mega-environments. By subdividing the environment into smaller homogeneous environments (megaenvironments), plant breeders can have more environment specific genotypes, thus presenting the opportunity of exploiting repeatable GxE across years (Gauch and Zobel, 1997, Yan et al., 2001). Alternatively, they can develop superior genotypes across a range of environments which show a high level of phenotypic stability and yield potential (Kanouni et al., 2015). For breeders to do this, they need to incorporate appropriate selection methods that integrate high yield potential and stability (Gauch et al., 1996). The genotype main effect and genotype by environment interaction (Staggenborg and Vanderlip, 2005) biplot allows for the selection of genotypes which are stable and have high yield potential, while addressing mega-environment differentiation at the same time, thus matching genotype performance with the megaenvironment (Yan et al., 2001). The GGE biplot simultaneously provides a visualisation of stability, mean performance and delineates the mega-environments providing plant breeders with a powerful analysis tool (Yan and Kang, 2002). The ideal genotypes in a biplot should exhibit high principal component (PC) 1 scores which denote high yields and low PC2 scores which represent high stability (Hamayoon et al., 2011). The ideal environment discriminates genotypes based on genetic differences as well as the target environment for which they are selected for (Gauch and Zobel, 1997). The genotype (G) and the GxE are the main sources of variation in the biplot genotype evaluation whereas the environment (E) is not relevant in biplot analysis (Yan and Tinker, 2006, Gauch and Zobel, 1997). The GGE biplot thus takes G and GxE into account and excludes the environmental and residual effect. Extensive research on genotype stability has been carried out using various stability measures (Lin et al., 1986). However, little attention has been paid to selecting chickpea genotypes that are both stable and have high yield potential. The hypothesis for this chapter was there was GxE in the evaluated chickpea genotypes and that the test environments were highly discriminating and representative. The objectives of this chapter were to: i) measure the extent of genotype by environment interaction in chickpea grown under varying environmental conditions, ii) explore the possibility of delineating the target environments into mega-environments, iii) identify stable chickpea genotypes with high yield potential and iv) identify the best performing genotypes for specific mega-environments. 73 6.2 Materials and methods A field experiment was conducted at the IA Watson research station at The University of Sydney in Narrabri, northwest New South Wales as described under the general Materials and Methods in Chapter 3. The total number of entries was 36 (as listed in Table 3.1 in Chapter 3) and they were planted in an alpha lattice design replicated twice each year and grown for two years. The entries were planted in a combination of no till and till, with and without irrigation. Every combination in each year was considered as a separate environment to give a total of eight environments (Table 6.1). Yield data were analysed using Genstat® edition 18 to determine the effect of season on yield variation. This was done by using REML analysis and fitting tillage, moisture regimes, genotypes and year in the fixed model, and range and row in the random model. Further analysis was performed to determine if there was GxE in the tested materials by using each tillage by moisture by year combination as a single environment. Genotype by environment was assigned to the treatment structure whereas the replicate was assigned to the blocking structure. GGE biplots were constructed using the GGE function in Genstat®. Weather data (temperature, rainfall and rainfall distribution) for each genotype at the vegetative, flowering and podding phenophases were computed and analysed to determine the means for each phenophase. Table 6.1: Field environments with different tillage and moisture regimes for analysis of chickpea phenotypic stability Environment Code Management regime 1 IRC14 Irrigation, under tillage, 2014 season 2 IRC15 Irrigation, under tillage, 2015 season 3 IRN14 Irrigation, under no till, 2014 season 4 IRN15 Irrigation, under no till, 2015 season 5 RFC14 Rainfed, under tillage, 2014 season 6 RFC15 Rainfed, under tillage, 2015 season 7 RFN14 Rainfed, under no till, 2014 season 8 RFN15 Rainfed, under no till, 2015 season 74 6.3 Results 6.3.1 Weather data There were significant differences (P<0.05) in all the weather parameters measured in 2014 and 2015, except for minimum temperature at the flowering phase (Table 6.2). Mean minimum temperature was lower in 2014 compared with 2015 with a similar trend observed in mean maximum temperature in 2014 compared to 2015. Rainfall was lower in 2014 with 74.7 mm recorded during the vegetative phase compared with 2015 which received almost double the amount (140.9 mm). Rainfall events were more common in 2015 with 28 rainy days recorded compared with 22 in 2014. Mean minimum temperature at flowering was similar in both years at 7.5°C, whereas the mean maximum temperature at the same phenophase was almost 5°C lower in 2014 (20.7°C) compared with 2015 (25.3°C). Rainfall was more than double, and rain events close to sixfold, in 2014 at the flowering phase compared with 2015 at a similar phenophase. The mean minimum temperature at the podding phase was more than 3°C lower in 2014 (11.3°C) than 2015 (14.6°C), with the same trend observed in the mean maximum temperatures of 25.3°C and 30.2°C in 2014 and 2015, respectively. Rainfall distribution was slightly better in 2015 than in 2014. In general, the 2015 season was hotter than the 2014 season and also had higher rainfall although it was poorly distributed. Table 6.2: Average weather conditions for each environment experienced by chickpea genotypes analysed for phenotypic stability Phenophase Vegetative Flowering Podding Weather IRN RFN IRC RFC IRN RFN IRC RFC 2014 2015 code 14 14 14 14 15 15 15 15 MnT 4.7 4.7 4.7 4.6 5.5 5.2 5.5 5.2 4.7a 5.3b MxT 17.2 17.2 17.2 17.2 19.7 19.3 19.7 19.2 17.2 a 19.4 b RF 79.1 73.7 75.4 70.6 141.8 140.2 141.4 140.2 74.7 a 140.9 b RD 23.4 22.3 22.5 21.6 28.5 28.0 28.4 28.0 22.4 a 28.2 b MnT 7.7 7.4 7.7 7.3 8.1 7.2 7.8 7.0 7.5 a 7.5 a MxT 21.0 20.5 20.9 20.3 25.9 25.0 25.5 24.6 20.7 a 25.3 b RF 78.6 50.0 81.9 53.1 43.0 7.6 46.7 10.5 65.9 a 27.0 b RD 12.5 13.2 13.5 13.9 2.9 1.5 3.2 2.0 13.2 a 2.4 b MnT 11.5 11.0 11.5 11.1 15.0 14.3 14.8 14.2 11.3 a 14.6 b MxT 25.9 24.8 25.7 25.0 30.1 30.3 30.2 30.4 25.3 a 30.2 b RF 47.3 11.0 47.1 9.9 46.6 17.1 43.2 16.1 28.8 a 30.7 b means means 75 RD 4.5 2.1 4.3 2.0 5.9 4.0 5.4 3.4 3.2 a 4.7 b MnT, mean minimum temperature; MxT, mean maximum temperature; RF, rainfall; RD, rainfall distribution. Environment codes are listed in Table 6.1. Means followed by a different letter in each row denotes that they are significantly different at P<0.05. 6.3.2 Grain yield under different environments Chickpea mean grain yield was higher in 2014 than in 2015 with different environments producing different yields. In general, no till environments had higher yields than till, and irrigated environments yielded more than rainfed. The highest yielding environment was IRN14 with a mean of 2148 kg ha-1 followed by IRC14 at 1999 kg ha-1 (Figure 6.1). RFC14 and RFN14 environments had very similar grain yields of 1712 kg ha-1 and 1714 kg ha-1, respectively. The environment with the lowest yield was RFC15 with a mean of 1145 kg ha-1. The environment yield ranking was; IRN14 > IRC14 > RFN14 > RFC14 > IRN15 > RFN15 > IRC15 > RFC15. There was more grain yield variability in 2014 than in 2015 with IRC15 and IRN15 environments resulting in very similar grain yield. 76 Figure 6.1: Chickpea grain yields for different tillage and moisture regime environments over two years (2014 and 2015). IRC14 and IRC 15 means irrigation + tillage in 2014 and 2015 season, respectively. IRN14 and IRN15 means irrigation + no till in 2014 and 2015 season, respectively. RFC14 and RFC15 stands for rainfed + tillage in 2014 and 2015 season, respectively. RFN14 and RFN15 stands for rainfed + no tillage in 2014 and 2015 season, respectively. 6.3.3 Factors accounting for grain yield variation Tillage (no till and till), moisture (irrigated and rainfed), genotype and year (2014 and 2015) main effects were significant at 95% confidence interval and explained a large proportion of the variation in the grain yields observed (Table 6.3). The genotype main effect explained 14.2% of the variation in grain yield and moisture levels explained 9.6%. Tillage had the lowest main effect factor explaining only 2.1% of the total variation in grain yield. The largest variation in grain yield was explained by the year (part of environment) main effect which accounted for 58% of the total variability. There was a significant genotype by year interaction 77 (P<0.05) which accounted for 5.5% of the total variation in grain yield. Moisture regime and year had a significant interaction and accounted for 2.3% variation, and the tillage by year effect was also significant and accounted for 0.2% variation in grain yield. There was a significant three-way interaction between tillage, moisture regime and year and it accounted for 0.6% of the total variation in yield. The remaining interactions were not significant and accounted for the remainder of the variation in yield. Table 6.3: The main factors accounting for grain yield variation in chickpea grown across different environments. Parameter TSS Percent of TSS Tillage 60.2 2.12*** Moisture 273.1 9.64*** Genotype 401.6 14.18*** Year 1641.7 57.97*** Tillage.Moisture 0.9 0.03 Tillage.Genotype 35.2 1.24 Moisture.Genotype 42.5 1.50 Tillage.Year 6.9 0.25*** Moisture.Year 64.7 2.28*** Genotype.Year 156.2 5.52*** Tillage.Moisture.Genotype 41.1 1.45 Tillage.Moisture.Year 15.7 0.55*** Tillage.Genotype.Year 23.9 0.84 Moisture.Genotype.Year 37.9 1.34 Tillage.Moisture.Genotype.Year 30.5 1.08 Values followed by an asterisk(s) indicate significant difference at *P<0.05, **P<0.01, ***P<0.001. 6.3.4 Genotype, environment and genotype by environment interaction There was a significant genotypic difference (P<0.001) among the genotypes tested which accounted for 12.6% of the total variation observed (Table 6.4). The test environments were significantly different and accounted for 66% of the total variation in grain yield. The interaction between genotype and environment was significant, indicating the genotypes had different yield rankings in the different environments. The genotype, and genotype by environment interaction accounted for cumulative variance of 24.6% of the total variation observed in grain yield. 78 Table 6.4: Combined analysis of variance (ANOVA) for genotype and environment effects on mean chickpea grain yields Source of variation d.f. s.s. m.s. v.r. F pr. % TSS Rep stratum 1 16641 16641 0.54 Genotype 35 11992315 342638 <0.001 12.6 Environment 7 62672976 8953282 290.37 <0.001 66.0 Genotype.Environment 245 11412558 46582 <0.001 12.0 Residual 286 8818513 Total 574 94909369 11.11 1.51 30834 9.3 Where d.f is degrees of freedom, s.s is sums of squares, m.s is mean sums of squares, v.r is variance ratio, Fpr is the Fischer test probability and TSS is total sums of squares. 6.3.5 Test environment evaluation The first PC accounted for 59% of the total variation in yield, PC2 accounted for 17% of the total variation and together, PC1 and PC2 accounted for 76% of the total variation in yield (Figure 6.2). The 2014 environments were positively correlated with each other with IRC14 and RFN14 showing higher similarity compared with RFC14 and IRN14. The highest dissimilarity between the 2014 environments was between IRC14 and RFC14 which is signified by the wider angle between their environmental vectors formed from the origin in Figure 6.2. The similarity in the 2015 environments was based on tillage practices rather than moisture regimes. IRN15 and RFN15 had a narrow angle between the two environmental vectors with a similar trend observable between IRC15 and RFC15 (Figure 6.2). The highest level of similarity in test environments in 2015 was between IRC15 and RFC15. There was a negative correlation between IRN15 and IRC14 because the angle between their environmental vectors was greater than 90° indicating a moderately large GEI. RFN15 and IRC14 had no relationship as evidenced by the 90° angle between their environmental vectors. 79 Figure 6.2: Environment scatter plot for evaluation of the test environment and chickpea genotypes. Red spheres indicate genotypes (see Table 6.1 for genotype codes) and green/blue arrows indicate environments. The larger the angle between two blue lines, the larger the difference between the test environments. 6.3.6 The ideal test environment The length of the environmental vector in the principal component is relative to the standard deviation of the particular environment and indicates the discriminating ability of that environment. The most discriminating environments for grain yield were IRC14 and IRN14 whereas the least discriminating were RFC14, RFC15 and IRC15 (Figure 6.3). Representative environments have small angles between them and the average environmental axis. The most representative environments were RFC15, IRC15 and RFC14 followed by IRN14 and RFN14. The least representative environments were IRN15 and RFN15 even though they were discriminating. An ideal environment should be both discriminating and representative. IRN14 was the ideal environment because it was both discriminating and representative and located 80 near the centre of the concentric circles in Figure 6.3. This environment was characterised by slightly higher and better rainfall distribution patterns at the vegetative phase compared to other environments in the 2014 season. However, a lower rainfall was recorded when compared to the 2015 season. IRN14 had a relatively high rainfall with good distribution at flowering and it was cooler than the 2015 environments. The other environments close to the ideal were RFN14 and RFC14 indicating that 2014 was generally a better growing season than 2015. IRN15 and RFN15 were discriminating but not representative so they may be useful for selecting specifically adapted genotypes. Figure 6.3: Test environment comparison scatter plot for evaluating genotype and environment interactions in chickpea yield. Red spheres indicate genotypes (see Table 6.1 for genotype codes) whereas green triangles indicate environments. The arrow in the middle of the concentric rings denotes the ideal environment and the further away a particular genotype is from the centre, the less ideal it is. The concentric rings helps one visualise how far a genotype is from the ideal environment. 81 6.3.7 Mean grain yield performance and stability test Genotypic stability is measured by the length of the perpendicular line to the average environment axis on either side and the proximity of the genotype to the average environment coordinate in a PC analysis biplot (Figure 6.4). The most stable genotypes were PBA Hattrick (6), Jimbour (19) and ICCV 98801 (15) however they were not the highest yielding. Sonali (9) was less stable compared with PBA Hattrick but had high yield potential, hence making it a good target for plant breeders (Figure 6.5). ICCV 96853 (14) and PBA Slasher (6) were less stable than PBA Hattrick but both out-yielded this genotype. Sonali exhibited higher GEI than both PBA Slasher and ICCV 96853 based on the length of the perpendicular line to the average environment axis. The most unstable genotype and low yielding genotype was ICCV 05308 (22) followed by ICCV 98813 (16). Genesis 079 (12) was unstable but had close to average yield across all the test environments. Amethyst (1), Lyle (26) and Sim (29) had very little contribution to both genotype and GxE since they clustered near the biplot origin. Both ICCV 05308 (22) and Genesis 079 (12) expressed high GxE and had low and average yield potential, respectively. 82 Figure 6.4: Principal component analysis scatter plot for evaluating grain yield performance and stability in chickpea genotypes. Red spheres indicate genotypes (see Table 3.1 for genotype codes) and green triangles indicate environments. 6.3.8 Selecting the ideal genotype The ideal genotype is that which is located in the middle of the concentric circles (Figure 6.5) and other genotypes near the centre of the concentric circles are considered equally as good. The best genotype, which had high yield potential and stability, was Sonali (9) (Figure 6.5). It was closely followed by PBA Slasher (7) and ICCV 96853 (14) with PBA Hattrick (6) slightly behind them. Genotypes ICCV 05308 (22), ICCV 98813 (16), Genesis Kalkee (3) and Genesis 079 (12) were distant from the ideal genotype making them less preferable for growing. Sonali had high yields in 2014 and the environments for this genotype could be ranked on yield as follows; IRN14 > IRC14 > RFN14 > RFC14. 83 Figure 6.5: Scatter plot for evaluating the ideal chickpea genotype. Red spheres indicate genotypes (see Table 6.1 for genotype codes) and green triangles indicate environments. 6.3.9 Mega-environment analysis The genotypes in the mega-environment (MGE) plot fell into seven sections delineated by the perpendicular lines from the origin and the environments fell into two sections (Figure 6.6). The eight environments were grouped into two MGE with IRC15, RFC15, RFC14, RFN14, IRN14 and IRC14 clustering into one mega-environment (MGE1) and IRN15 and RFN15 in the other mega-environment (MGE2). The till regimes (RFC14, RFC15, IRC14, IRC15) clustered in the same mega-environment in both test years indicating repeatability of results under tillage. The vertex genotypes which were located the furthest in each sector were joined using equality lines to form a polygon such that all the other genotypes were inside the polygon (Yan and Tinker, 2006). These vertex genotypes were the most responsive in each section of 84 the plot. Sonali was the vertex genotype in the MGE1 cluster, therefore the best performer, closely followed by PBA Slasher and ICCV 96853 which were above average performers. The equality line in MGE1 connects Sonali (9), ICCV 96853 (14), PBA Slasher (7) and Jimbour (19) so the performance ranking for MGE1 is as follows; Sonali > ICCV 96853 > PBA Slasher > Jimbour. There were no specific high yielding genotypes in MGE2 since it lacked a vertex genotype. Genesis Kalkee (3), ICCV 05308 (22) and ICCV 98813 (16) were below average performers in all the test environments. Figure 6.6: Mega environment scatter plot for evaluating chickpea yield across environments. Red spheres indicate genotypes (see Table 3.1 for genotype codes) and green triangles indicate environments. The green line represents the equality line for joining vertex genotypes. 6.4 Discussion Rainfall is the most important factor affecting crop production in rainfed agriculture (Godwin, 1990). Rain distribution plays a key role as well in explaining variation in yield. Gangopadhyaya and Sarker (1965) reported about 75% of the total variation in maize yield was 85 accounted for by rainfall distribution. In the present study, rainfall and rainfall distribution between 2014 and 2015 caused significant differences at various chickpea phenophases. There was less rainfall during the vegetative phase in 2014 than in 2015, however the yields were higher in 2014 indicating that this may not be a critical stage for yield requiring high moisture levels. There was a large difference in rainfall and its distribution during the flowering phase in 2014 compared to 2015 with 2014 receiving more and better distributed rainfall. This may have contributed substantially to the high yields that were observed in 2014 compared with 2015 as the reproductive phase is the most sensitive to water stress (Nayyar et al., 2006, Mafakheri et al., 2010). The year effect was the largest contributor to the variation in yield observed between the two years with similar findings reported for chickpea in northern New South Wales in Australia (Haigh et al., 2005). The difference observed in rainfall and its distribution at the podding stage in 2014 and 2015 was not as large as that observed during the flowering stage. Temperature plays a key role in determining chickpea yield with temperatures less than 10°C (Chaturvedi et al. (2009) and more than 30°C (Summerfield et al. (1984) causing a reduction in grain yield. Minimum mean temperature differences were not large during the vegetative and flowering phases in 2014 and 2015. Notably, 2015 was warmer than 2014 by 5°C in terms of mean maximum temperature at both flowering and podding stages with the podding stage exposed to temperatures greater than 30°C in 2015. This may have contributed to the lower yields recorded in 2015. The ideal test environment in terms of discriminating ability and representativeness was IRN14 followed by RFN14. Both were no till environments suggesting that no till may be beneficial in farming systems through improved soil aggregate stability (Chan and Mead, 1988), improved soil structure (Page et al., 2013) and increased water storage (Radford et al., 1995). No till environments (IRN14, RFN14, IRN15 and RFN15) were the most discriminating compared with till environments (RFC14, RFC15 and IRC15) except for IRC14 which was as discriminating as the no till environments. Genotype stability is only effective if it is accompanied by high yields. The most desirable genotypes are those which show high stability and have high yield potential. In the GGE biplot, genotypes with a high PC1 score are high yielding and a low PC2 score are stable (Maqbool et al., 2015). In the present study, PBA Slasher and ICCV 96853 had both high stability and high yield potential. Sonali was less stable than PBA Slasher and ICCV 96853 but out-yielded both those genotypes. Genotypes that have low stability but high yield potential may be selected for 86 a specific environment, whereas genotypes that have low yield and high stability are the most undesirable in a breeding program (Yan and Wu, 2008) Extensive breeding programs strive to save costs by reducing the number of testing sites. It is therefore important to identify testing sites which are highly discriminating and at the same time representative. This allows the plant breeder to effectively reduce the number of test sites and costs while at the same time managing the selection process effectively such that no desirable genotypes are discarded (Imtiaz et al., 2013). In the present study, two megaenvironments were identified; MGE1 comprising of IRC15, RFC15, RFC14, RFN14, IRN14 and IRC14, and MGE2 comprising of IRN15 and RFN15. The two MGE accounted for 76% of the total GEI which was explained by the genotype-mega environment variance component. IRN14 and RFN14 represent the rest of MGE1 as good test environments and either IRN15 or RFN15 would suffice for MGE2. This can reduce the test sites from eight to three, which is a cost effective strategy. By identifying the MGE, selection based on individual MGE can be done so that the best adapted genotypes are cultivated (Yan et al., 2001). In the present study, the best adapted genotypes for MGE1 are Sonali, PBA Slasher and ICCV 96853. There is no specifically adapted genotype for MGE2. 6.5 Conclusions The analysis presented in this chapter revealed that there was significant GxE in chickpea grown under varying environmental conditions. The year effect was the largest contributor to the observed variation in yield and this was driven by the weather conditions in each season. Rainfall and rain distribution played a key role in yield formation in the test environments with seasons that had high rainfall which was well distributed yielding better than others. The GGE biplot is an effective tool in selecting good test environments, ideal genotypes and assessing the possibility of grouping the environment into mega-environments. In the present study, GGE successfully grouped the environment into two MGE and ideal genotypes relative to the performance of the other genotypes in similar environmental conditions were identified. Sonali, PBA Slasher and ICCV 96853 were identified as suitable genotypes for cultivation under rainfed or irrigated regimes with no till or tillage, hence showing a wider adaptation. 87 CHAPTER 7: DEVELOPMENT OF A DROUGHT TOLERANT CHICKPEA IDEOTYPE FOR THE AUSTRALIAN GRAIN BELT 7.1 Introduction Plant breeders have selected for yield empirically over the years (Donald, 1968) based on genetic variation. This variation is caused by mutation, recombination of genes during reproduction and lateral gene transfer. Plant breeders use this variation and knowledge of gene, environment and management interactions to develop high yielding crop cultivars. Further yield increases are achieved through conservation of soil moisture, control of pests and diseases and the use of fertilisers (Johnson, 1984). Breeding programs traditionally select the highest yielding genotypes in any given environment and cross these to generate high yielding progenies for advancement. The challenge with this approach is that very little is known about the physiological, morphological and biochemical drivers of yield in different genotypes in different environments. Furthermore, the heritability of yield is generally low (Ludlow and Muchow, 1990) because the expression of this polygenic trait is significantly influenced by the environment, including drought, thus reducing the repeatability of results (Johnson and Geadelmann, 1989). Under drought, secondary traits linked to yield which exhibit higher heritability than yield could be selected (Ludlow and Muchow, 1990, Blum, 1988). This ideotype approach is an alternative strategy to empirical breeding (Peng et al., 1994) and allows the breeder to predict the ideal genotype in the target environment. An ideotype is a biological plant model which behaves in a known manner when exposed to a distinct environment (Donald, 1968). Donald’s concept was to consolidate several important traits that may manifest in different genotypes into one ideal genotype that would perform better than the individual parents. Definition of the plant type (Rasmusson, 1987) provides plant breeders with clear cut objectives based on defined traits (Rasmusson, 1991) that provide a blueprint for pyramiding traits (Mock and Pearce, 1975). Thus ideotype breeding is more analytical than traditional empirical selection and breeding. One of the most important steps in ideotype breeding is the identification of the target environments (Mock and Pearce, 1975, Trethowan, 2014) and the target ideotype should perform optimally in these environments. Some of the key factors to consider in the target environment include temperature, soil moisture and soil fertility (Mock and Pearce, 1975). 88 Each ideotype is normally designed for a certain target environment and could possibly be grown in areas which lie in the same environmental type or mega-environment. The next step is identification of the physiological and morphological traits that contribute to yield either directly or indirectly. These traits should show genetic diversity to be incorporated into an ideotype breeding program (Rasmusson, 1987). Ideally the target traits should be easy to measure and highly heritable, however this should not preclude traits that are laborious to measure if they are important and correlated with yield (Rasmusson, 1987). Trait relationships must also be carefully considered because pleiotropy, trait compensation and inferior donor germplasm may influence the target ideotype thus reducing breeding progress (Rasmusson, 1991). The identified traits can then be pyramided in one genotype (Mock and Pearce, 1975). Crop modelling has recently become an important enabling tool in plant breeding (Tardieu, 2003, Hammer et al., 2006). From a modelling perspective, an ideotype is a set of defined crop parameters that drive growth and development in defined environmental conditions (Rotter et al., 2015). High quality long-term data is an imperative for model calibration and the generation of accurate simulation results (Rotter et al., 2015). These ideotype models can also be refined to capture variability in the climate (Rotter et al., 2015). Models have been a powerful tool in ideotype design and testing in silico (Semenov and Stratonovitch, 2013). Data on multiple sites over many years can be produced without running actual field trials, which reduces the cost of plant breeding. Chapman et al. (2002) emphasised that models provide an understanding of the temporal and spatial environmental effects on crops, especially when experimentation is not possible. Crop modelling can also be used to assess crop responses to environmental factors (White et al., 2002). Several crop ideotypes have been developed including rice (Khush, 1995) and wheat (Semenov and Stratonovitch, 2013). The software tool APSIM simulates cropping systems using climate, soil, management and crop genetic coefficients to predict the economic yield (Keating et al., 2003). The APSIM model uses the supply and demand concept of important plant growth resources (light, water, nitrogen and carbon) to create a plant phenotype (Hammer et al., 2001). There has been no attempt to develop and model the performance of a drought tolerant chickpea ideotype from a defined chickpea germplasm gene pool and compare the ideotype performance with drought tolerant and drought susceptible chickpea genotypes under different management practices (no tillage and full tillage systems) in the Australian grain belt. 89 This chapter aims to; i) develop a chickpea ideotype, ii) characterise the chickpea growing environments based on soil moisture deficits at various growth stages across the Australian grain belt, iii) identify the critical stages where drought occurs to better match phenology to environment, and iv) assess the performance of selected chickpea cultivars and a target ideotype across the Australian grain belt. 7.2 Materials and methods 7.2.1 Field experiments Data from an experiment conducted at The University of Sydney’s IA Watson Grains Research Centre at Narrabri (latitude 30.275616° S and longitude 149.803547° E) in 2014 and 2015 as described under Materials and Methods in Chapter 3 were used to develop the chickpea ideotype, and parameterise and validate the APSIM-Chickpea model (Version 7.8). The larger experiment comprised 30 entries (25 desi and 5 kabuli types – refer to Table 3.1) and for the purpose of the ideotyping presented in this chapter, five desi genotypes were chosen; Amethyst, Kyabra, PBA Hattrick, Tyson and Sonali. These genotypes were selected because of their differential response to drought based on yield rankings in well-watered and water stress conditions, as well as stress tolerance index. For example, Sonali has a high yield and is drought tolerant with a field stress tolerance index of 1.27 calculated according to (Fernandez, 1992). Tyson is reported to be drought tolerant (Sarma et al., 2011) but showed moderate tolerance in these field experiments with a stress tolerance index of 0.95. Tyson has also been used as a parent in breeding programs to develop new varieties (Lake et al., 2016). Amethyst had a stress tolerance index of 0.84 and was classified as drought susceptible based on the field evaluation at Narrabri. PBA Hattrick, widely cultivated by farmers in northern NSW had a stress tolerance index of 1.02 whereas Kyabra had a stress tolerance index of 0.87. Hence, PBA Hattrick had moderate drought tolerance and Kyabra was drought susceptible. 7.2.2 Chickpea ideotype development The chickpea ideotype was designed following the recommendations of Rasmusson (1987), Martre et al. (2015) and Rotter et al. (2015). Field data obtained from the two seasons (2014 and 2015) in Narrabri was used to construct the chickpea ideotype. The data were subjected to analysis using Genstat® edition 18 to generate means, test genetic variation of traits at 95% confidence levels, and generate least significant differences (LSD) at P<0.05 using the linear mixed models in the REML function (Patterson and Thompson, 1971). Multiple linear regression analysis was subsequently used to identify traits that significantly explained yield 90 variation. The measured traits were also subjected to correlation analysis following the method described by Snedecor and Cochran (1987). Trait relationships with yield and individual interrelationships were considered and traits optimised to a maximum (Marinho et al., 2014) or minimum depending on their correlation with yield (Figure 7.1). The maximum and minimum values were chosen as relevant and the LSD used to establish a range for the trait (Figure 7.1). The use of trait ranges was intended to give breeders some flexibility while targeting traits. The optimised values were generated and assigned to the ideotype and then subjected to analysis (Laurila et al., 2012) using Genstat® edition 18. Field Field experiments to generate data on traits for analysis Is there genetic variation? No Yes Select trait Discard trait No Discard trait Select trait No Select trait -ve +ve Trait relationship with yield High - LSD Low + LSD Low + LSD Does trait have high heritability? Yes Low priority -ve Does trait significantly explain yield variation? Yes Trait interrelationships +ve Average -ve LSD +ve Average LSD High - LSD Figure 7.1: Flow diagram for chickpea ideotype construction. LSD, least significant difference at P<0.05; +ve, positive correlation; –ve, negative correlation. 7.2.3 Environmental characterisation: the soil water deficit approach Soil water deficit at important phenophases (e.g. flowering, grain filling) of chickpea were estimated using the APSIM-Chickpea model. Fifty locations within the Australian grain belt based on the chickpea National Variety Trials (NVT) sites (http://www.nvtonline.com.au/) were selected. The critical chickpea planting window for each location was obtained from the chickpea sowing guides provided by each state. For each location, the SoilMapp iPad® application developed by Commonwealth Scientific and Industrial Research Organisation 91 (CSIRO) was used to search the Australian Soil Resource Information System (ASRIS) map discovery database to obtain the relevant soils. Once the soil type was identified for each location, it was selected from the APSIM soils repository and used for analysis. The SILO climate database for the period from 1905 to 2004 was used. Sowing was conducted when the cumulative rainfall in three consecutive days was greater than 15 mm within the sowing window. The sowing rule for each genotype was set to begin when there was at least 200 mm of allowable available soil water. The planting density for each genotype was set at 25 plants per m2 with a 30 mm depth and an inter-row spacing of 300 mm. The initial surface residue in the no-till management system was set at 1000 kg ha-1 with a 0.6 fraction of standing residue remaining. No fertilisers were applied to either no-till and till systems. In the historical analysis, 100 year weather, soil type and sowing rules for each location were considered. Simulations (60 000) were performed denoting combinations of six genotypes (including the ideotype), two management systems (till and no-till), 100 years and 50 locations. The crop growing season was divided into five stages; juvenile, floral initiation, flowering, start of grain filling, end of grain filling and maturity. The soil water deficit was derived from the water supply/water demand ratio (Lake et al., 2016, Kholová et al., 2013) and used to analyse soil moisture stress levels at each growth stage. The soil water deficit values were then subjected to average link cluster analysis using Euclidean distances in Genstat® edition 18. The output was used to identify various stress environments in the Australian grain belt. An index of 1 indicated no moisture stress (no drought) and 0 very low moisture (severe drought) as described by Kholová et al. (2013), with an index less than 0.7 considered a drought event (Lake et al., 2016). 7.3 Results 7.3.1 Chickpea ideotype A total of 21 parameters were identified through multiple linear regression on yield and they accounted for 91% of the total variation at P<0.001 (Table 7.1). High yielding genotypes developed ground cover early in the season and also had high NDVI both at early and late podding stages. These genotypes also produced their first flower earlier in the season and finished flowering earlier. Genotypes that continued flowering and had a late date to last flower were associated with low NDVI at flowering and low shoot harvest index, hence rendering NDVI an undesirable trait. However, a longer flowering period was associated with longer leaves and higher yield. Plant height had no clear correlation with yield but shorter machine harvestable plants were preferred because they were associated with early flowering, longer flower duration and high NDVI scores which all contributed positively to yield. Genotypes 92 with high chlorophyll content at mid-podding and low chlorophyll content at late-podding produced higher yield. Leaf characteristics (number of leaflets per leaf, single leaflet area and single leaf area) were not significantly correlated with yield. Average-sized leaves should be selected as opposed to small leaves because larger leaf area was associated with early flowering, longer flowering duration and longer leaves, all of which are desirable for increasing yield. Average shoot harvest index was selected because plants with high shoot harvest index also had slower development of ground cover, lower chlorophyll content at mid podding and lower NDVI. On the other hand, they flowered earlier and were shorter. All these traits influence yield in opposite directions, hence the average was chosen. Similarly, plants with high pod harvest index also had high NDVI. However, they had short narrow leaves, were taller, flowered later and had lower seed mass. Table 7.1: Wald statistic, correlations and decisions used to construct the chickpea ideotype Wald Parameter statistic Correlation Decisions Early ground cover (%) 14.2** 0.10 H - LSD Days to first flower 31.66*** -0.18 L + LSD Days to 50% flowering 18.63*** -0.23 L + LSD Days to last flower 24.69*** -0.36* L + LSD Flower duration (days) 24.62*** 0.12 H - LSD NDVI at early podding 58.92*** 0.55 H - LSD NDVI at late podding 8.32* -0.52* L + LSD Plant height at late flowering (Krupinsky et al.) 20.72*** 0.10 L + LSD Plant height at late podding (Krupinsky et al.) 29.84*** 0.04 L + LSD Chlorophyll content at mid podding (SU) 5.38* 0.02 H - LSD Chlorophyll content at late podding (SU) 9.61** -0.35* L + LSD Number of leaflets per leaf 13.42** -0.10 AV ± LSD Single leaf area (cm2) 25.64*** -0.04 AV ± LSD Single leaflet area (cm2) 35.79*** -0.03 AV ± LSD Leaflet length (cm2) 28.16*** 0.02 AV ± LSD Number of pods per plant 11.43** -0.18 AV ± LSD Pod biomass per plant (g) 19.11*** -0.31 L + LSD Pod harvest index 9.6** 0.10 AV ± LSD Shoot biomass 25.85*** -0.41* L + LSD 93 Shoot harvest index 47.88*** 0.24 AV ± LSD 1000 seed weight (g) 9.05** -0.39 L + LSD Wald statistic is from the multiple linear regression Wald tests for dropping terms. Significant terms at various levels of confidence were picked; * = P<0.05, ** = P<0.01, *** = P<0.001. Correlations are between the considered traits and yield. Decision on ideotype optimisation based on trait ranges and relationship with yield and other traits. L, low; H, high; AV, average; LSD, least significant difference at 5%; SU, SPAD Units. The early ground cover ranged from 13.2 – 25.9% with the selected ideotype at 23.2% (Table 7.2). Days to first flower, 50% flowering and days to last flower ranged from 68 – 93, 80 – 103 and 118 – 124, respectively, with the ideotype classified as 70, 82 and 119 days in the same trait order. Sonali was closest to the ideotype in days to flowering whereas the other genotypes flowered later. All genotypes stopped flowering around the same time regardless of when they started to flower. Flower duration ranged from 31 – 50 days with the ideotype classified at 48 days. NDVI at early podding ranged from 0.66 – 0.74 and 0.22 – 0.41 at late podding. Both NDVI and chlorophyll measures at the podding stages of the five selected genotypes were comparable to the ideotype. Variation in number of leaflets was low compared to leaf area. Number of pods per plant ranged from 21 – 54, and pod biomass ranged from 6.4 – 31.6 g. Shoot biomass varied greatly (13.6 – 52.6 g), however the range in shoot harvest index was relatively narrow (0.33 – 0.48). The ideotype values for shoot biomass and shoot harvest index were 17.7 g and 0.39, respectively. Seed weight also varied greatly with an observed range of 139 – 392 g and the ideotype was classified as 148 g. Table 7.2: Trait range, genotype and ideotype values for evaluating chickpea drought tolerance through APSIM modelling PBA Parameter Trait range Amethyst Kyabra Hattrick Sonali Tyson Ideotype Early ground cover (%) 13.2-25.9 15.7 22.2 18.9 15.1 16.2 23.2 Days to first flower 68-93 88 88 87 77 86 70 Days to 50% flowering 80-103 99 99 98 88 97 82 Days to last flower 118-124 120 122 121 118 120 119 Flower duration (days) 31-50 32 34 34 41 34 48 NDVI at early podding 0.66-0.74 0.70 0.71 0.72 0.69 0.70 0.72 NDVI at late podding 0.22-0.41 0.28 0.28 0.29 0.22 0.25 0.25 94 Plant height at late flowering 49-64 58 64 64 56 54 51 Plant height at late podding 54-69 62 68 69 60 56 55 Chlorophyll at mid pod (SU) 65-74 69 74 71 65 66 72 Chlorophyll at late pod (SU) 49-68 52 54 56 49 52 51 Number of leaflets per leaf 12-15 14 14 12 14 14 14 Single leaf area (cm2) 4-21 6 7 6 8 5 7 Single leaflet area (cm2) 0.3-1.6 0.4 0.5 0.5 0.6 0.3 0.5 Leaflet length (cm2) 1.0-2.2 1.2 1.2 1.2 1.3 1.0 1.2 Number of pods per plant 21-54 42 30 34 32 35 31 Pod biomass per plant (g) 6.4-31.6 12.4 10.5 10.0 10.5 8.5 8.8 Pod harvest index 0.75-0.82 0.81 0.81 0.80 0.79 0.81 0.80 Shoot biomass 13.6-52.6 25.0 24.8 21.6 18.1 16.2 17.7 Shoot harvest index 0.33-0.48 0.42 0.34 0.37 0.48 0.43 0.39 1000 seed weight (g) 139-392 224 206 192 139 148 151 SU, SPAD units All five selected genotypes were compared to the ideotype based on their performance against the 21 parameters used for ideotype construction. Sonali was the closest to the ideotype with 76% resemblance (Figure 7.2). This resemblance was primarily based on phenology (days to first flower, days to 50% flowering, days to last flower and flower duration), leaf characteristics and number of pods per plant. The next closest resemblance to the ideotype was Tyson at 73.2% similarity. This resemblance was based on pod biomass, NDVI, plant height and chlorophyll content. Kyabra and PBA Hattrick were the most closely related pair of genotypes with 90.3% similarity, followed by Tyson and Amethyst with 83.7% similarity. Sonali and Tyson showed drought tolerance under field conditions (Narrabri) with stress tolerance indices of 1.27 and 0.95, respectively, compared to the drought susceptible cultivar Amethyst (0.84). 95 90.3 76.0 73.2 79.8 83.7 Figure 7.2: Evaluation of chickpea for drought tolerance using minimum spanning tree for genotype similarity. Genotypes close to each other along the line are more similar than those further away in the tree. The x-axis is dimension 1 and the y-axis is dimension 2 of the Genstat output. 7.2 Validation of the APSIM-Chickpea model The simulated days to flowering compared to the observed days to flowering in the Narrabri field experiment returned a coefficient of determination of 0.6 and a root mean square error of 12 (Figure 7.3a). When the 1:1 line was fitted, it showed that the simulated values were slightly underestimated. The coefficient of determination for simulated and observed yield was 0.7 with a root mean square error value of 823 (Figure 7.3b), thus the simulated yield was slightly overestimated. a b 110 4000 r 2 = 0 .6 Y = 0 .4 8 0 3 * X + 3 8 .4 R M SE = 12 S im u la te d S im u la te d 100 90 3000 2000 80 r 2 = 0 .7 Y = 0 .8 1 * X + 1 1 0 3 1000 70 70 80 90 O b served 100 110 R M SE = 823 1000 2000 3000 4000 O b served Figure 7.3: Evaluation of chickpea traits using APSIM modelling. (a) Days to 50% flowering 96 (days) and (b) grain yield (kg ha-1) for observed and simulated data. R2 is the coefficient of determination and RMSE is the root mean square error. 7.3 Simulated yield The simulated yield from the 50 locations ranged from 760 to 3902 kg ha-1 showing the diversity of production environments investigated (Figure 7.4a). No till environments had a slightly higher average yield of 2559 kg ha-1 compared to 2492 kg ha-1 under till (Figure 7.4b). The chickpea ideotype had the highest average yield of 2678 kg ha-1 compared to Sonali, PBA Hattrick, Amethyst, Kyabra and Tyson which yielded 2553, 2513, 2487, 2487 and 2457 kg ha1 , respectively (Figure 7.4b). a 4000 3000 3000 K g ha -1 kg ha b -1 4000 2000 1000 2000 1000 0 0 L o c a tio n s in a s c e n d in g o r d e r AM ID K Y PH SO TY Figure 7.4: Evaluation of chickpea yields across different production environments (a) Simulated yield (kg ha-1) range representing 50 locations from the lowest yielding location to the highest (b) Actual yield (kg ha-1) of individual genotypes under no-till (checked) and till (diagonal lines) environments. AM, Amethyst: ID, Ideotype: KY, Kyabra: PH, PBA Hattrick; SO, Sonali; TY, Tyson. 7.4 Environmental characterisation and soil water deficit patterns Cluster analysis performed on the soil water deficit output from APSIM grouped the environments into three major clusters and two ungrouped environments at 95% similarity value (Figure 7.5 and Table 7.3). The first cluster was comprised of four sites only; Albany, Hamilton, Minlaton, Riverton. The second cluster represented 33 locations making it the largest group, and the third cluster comprised 10 locations. The two ungrouped locations were Bourke and Rudall. The complete cluster groups and names are listed in Table 7.3. 97 Figure 7.5: Dendrogram of Australian chickpea production environment characterisation based on soil moisture deficit. Arrows indicate the start of a new cluster or group. Table 7.3: Australian chickpea production environmental clusters based on soil water deficit Cluster 1 Number Location 1 Albany 2 Hamilton 3 Minlaton 4 Riverton Cluster 2 Number Location 1 Bellata 2 Biloela 3 Birchip Cluster 3 Number Location 1 Capella 2 Emerald 3 Griffith 4 Hillston 5 Merredin 6 Mullewa 7 Mungindi 8 Rainbow 9 St George 10 Walgett 98 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 Brookstead Carnamah Coonamble Dalby Dubbo Edgeroi Forbes Geraldton Goondiwindi Hermitage Horsham Kaniva Mingenew Moree Mundulla Narrabri North Star Northam Roma Rutherglen Springsure Tamworth Trangie Tulloona Wagga Wagga Warren Warwick Wongan Hills Yeelana York Ungrouped 1 Rudall 2 Bourke In the first cluster, genotypes flowered and matured on average at 89 and 175 days, respectively (Figure 7.6a). The second cluster flowered slightly earlier at 86 days after sowing and matured earlier at 159 days. The third cluster was the earliest flowering and maturing of the three clusters at 79 and 147 days, respectively. The ungrouped locations flowered at 70 days and matured at 128 days on average. Cluster 1 was the highest yielding (3645 kg ha-1) followed by clusters 2 (2700 kg ha-1) and 3 (1801 kg ha-1) (Figure 7.6b). The two ungrouped locations had a low mean yield of 928 kg ha-1. 99 a 200 3000 Y ie ld 150 N o. of days b 4000 100 2000 1000 50 0 0 1 2 3 C lu s te r s 4 1 2 3 4 C lu s te r s Figure 7.6: Evaluation of APSIM-predicted chickpea traits for drought tolerance based on soil water deficit clusters. (a) Mean days to 50% flowering (grey bars) and maturity (black bars) and (b) mean grain yield (kg ha-1). Numbers on the x-axis represent the cluster numbers. The first cluster had a 97% chance of yielding over 2500 kg ha-1 in every season and less than a 0.3% chance of yielding less than 1000 kg ha-1 (Figure 7.7). Cluster 2 had a 60% chance of exceeding 2500 kg ha-1 per year and an approximately 30% chance of yielding between 10002500 kg ha-1. The third cluster had a 30% chance of yielding 2500 kg ha-1 or more with an equal chance of yielding 1000 kg ha-1 or below. The two ungrouped locations had a mean yield of 928 kg ha-1. All the simulated genotypes produced similar yield patterns with little variation in the frequencies in all the clusters with the ideotype performing better in all comparisons (Figure 7.7). 100 A m e th y st I d e o ty p e K yabra P B A H a ttr ic k S o n a li T y so n 100 C lu s te r 1 80 60 40 0 C lu s te r 2 80 60 40 20 0 80 C lu s te r 3 F re q u e n cy (% ) 20 60 40 20 U ngrouped 0 80 60 40 20 0 0 50 >2 -2 00 10 <1 00 50 0 0 0 50 >2 -2 00 10 <1 00 50 0 0 0 50 >2 0 50 -2 00 10 <1 00 0 0 50 >2 -2 00 10 <1 00 50 0 0 0 50 >2 -2 00 10 <1 00 50 0 0 50 >2 -2 00 10 <1 00 50 0 0 0 -1 Y ie ld (k g h a ) Figure 7.7: Frequency predictions (%) for chickpea yield based on cluster groupings (identified in Figure 7.5). 7.5 Stress timing and the critical period for yield penalty There was adequate soil moisture on average at all locations during the juvenile development stages (Figure 7.8). However, a gradual decline in soil moisture occurred from the juvenile stage to flowering in all three clusters. The two ungrouped locations (Bourke and Rudall) experienced a sharp moisture decline immediately after the juvenile stage. Cluster 1 maintained soil moisture all the way to maturity with a gentle decline during the grain filling period, levelling off at maturity. Cluster 2 and 3 experienced a sharp moisture decline from flowering until the end of grain filling. While terminal drought was experienced in both clusters 2 and 3, the intensity of drought was greater in cluster 3. The two ungrouped locations experienced both intermittent and terminal drought. 101 1 .5 S o il w a te r d e fic it C lu s te r 1 C lu s te r 2 1 .0 C lu s te r 3 U n grou p ed 0 .5 0 .0 JV FI FL SGF EG F MT G r o w th sta g e Figure 7.8: APSIM-predicted soil water deficits in different growth stages over a 100 year period x 50 locations x six varieties. JV, juvenile stage; FI, floral initiation; FL, flowering; SGF, start of grain filling; EGF, end of grain filling; MT, maturity stage. Multiple linear regression of all growth stages was significant (P<0.001) and accounted for 96.2% of the total variation in grain yield. The start of grain filling was the critical point where drought most severely affected yield (Table 7.4). However, stress later in grain-filling also limited yield but to a lesser extent. These results show that the whole grain filling period is very sensitive to any soil water deficit. Table 7.4: Multiple linear regression of various growth stages in relation to chickpea yield Summary of analysis Source d.f. s.s. m.s. v.r. F pr. Regression 2 22550219 11275109 619.05 <.001 Residual 47 856032 18213 Total 49 23406251 477679 Wald tests for dropping terms Term Wald statistic d.f. F statistic F pr. swdSGF 493.6 1 493.64 <0.001 swdEGF 10.8 1 10.8 0.002 Where swdSGF is soil water deficit at the start of grain filling and swdEGF is soil water deficit at the end of grain filling, d.f is degrees of freedom, s.s is sums of squares, m.s is mean sums of squares, v.r is variance ratio, Fpr is the Fischer test probability 102 7.6 Discussion A target crop ideotype, as defined by Donald (1968), is a developed biological model with predictable behaviour in a known environment. Defining the target environment constitutes an important step in ideotype breeding (Trethowan, 2014). The ideotype developed in the present study is a stress ideotype (Sedgley et al., 1990) suited for areas which experience terminal drought. One of the proposed traits for this ideotype is early ground cover (Singh et al. (1993); Toker et al. (2007). This enables the plant to cover the bare ground quickly, thereby reducing water loss associated with evaporation from the soil (Sekhon et al., 2010). Early flowering is important especially in areas which experience terminal drought because the plant can complete pod setting before the onset of water stress (Jain, 1975: Singh et al., 1993: Toker et al., 2007). However, there is often a trade-off between early flowering and high yield potential which may limit yield in cooler, wetter years. Short machine harvestable plants are desirable in water limited environments partly because they will not waste resources in the stem and are less susceptible to lodging. Small leaflets which reduce water loss through evapotranspiration are desirable and adopted for the ideotype designed in the present study. Similar proposals were made by Toker et al. (2007), Saxena (2003) and Saxena and Johansen (1990). High chlorophyll content at mid podding was associated with higher yields under drought conditions whereas high chlorophyll later in the growing season resulted in lower yields (Jain, 1975; Nayyar et al., 2005). This is probably a function of slightly later development, thus exposing the plant to moisture stress during late pod filling. High harvest index is an important determinant of yield under moisture limited conditions (Siddique and Sedgley (1985) Jain (1975). Sonali and Tyson were closer to the ideotype than the widely grown cultivar PBA Hattrick in terms of harvest index. APSIM is a dynamic crop simulation model that takes into account management options in farming systems to simulate both biological and physical processes (Keating et al., 2003). It has been effectively parameterised for various crops including mungbean, peanut and chickpea (Robertson et al., 2002), wheat and soybean (Mohanty et al., 2012), pearl millet (Akponikpè et al., 2010), sorghum (Whish et al., 2005) and maize (Archontoulis et al., 2014). In the present study, the comparison between simulated and observed field data returned a coefficient of determination for days to flowering of 0.6 and 0.7 for yield. These data are comparable to Carberry (1996) and Robertson et al. (2002) who each reported a coefficient of determination of 0.7 for days to flowering and yield in chickpea, respectively. 103 Grain yield ranged from 760 to 3902 kg ha-1 in the Australian grain belt environments and similar diversity was reported by Chauhan et al. (2008). These authors reported six clusters of Australian environmentally-based locations compared to just three in India. This diversity reflects the importance of yield stability for both plant breeders and grain growers. However, this challenge can be tackled by exploiting genotype by environment by management interactions and matching crop phenology to the target environment. The majority of the locations had simulated yields greater than the break-even yield for chickpea of 1 t/ha reported by Whish et al. (2007) which makes chickpea a profitable venture for farmers. High yield in some locations, coupled with the benefits of soil amelioration that chickpea provides, should lead to wider adoption of chickpea in the Australian grain belt farming systems. The yield in the no-till production systems was consistently higher than the till system as observed by others (Dalal, 1989, Horn et al., 1996). This advantage is perhaps due to water conserved in the notill system that becomes available later in the growing season (Rathore et al., 1998). Kholová et al. (2013) used the soil moisture deficit approach to characterise sorghum growing environments. Lake et al. (2016) and Chenu et al. (2011) used the same approach to characterise chickpea and wheat growing environments, respectively. The present study grouped environments into three distinct clusters with two arid locations (Bourke, Rudall) remaining ungrouped. This classification differs slightly from Lake et al. (2016) who reported four clusters. Nevertheless, the stress patterns are similar to the Lake study with the majority of locations classified as limited by terminal drought. However, in the current study no environments recovered from terminal drought as reported by Lake et al. (2016). Stress generally started at the reproductive phase, with early podding/start of grain filling being the most sensitive to drought. A similar finding was reported by (Thudi et al., 2014). 7.7 Conclusions Ideotype breeding can increase chickpea drought tolerance and hence sustain yields across the Australian grain belt and areas with similar climates. In silico testing is a more efficient way to evaluate chickpea genotype performance in a wide range of environments. The developed chickpea ideotype outperformed the other genotypes in a wide range of environments and was closely followed by Sonali which was identified as a drought tolerant genotype. Since Sonali had 76% similarity to the ideotype, it can be used as a target for incorporating the ideotype traits. Incorporating traits associated with drought tolerance into commercially grown genotypes can lead to faster adoption of drought tolerant genotypes and resilient chickpea production systems. 104 Based on the soil water deficit approach the Australian chickpea growing environment was characterised into three main clusters. The same approach can be used to characterise the growing environments for any crops grown in the Australian grain belt as well as other parts of the world where drought is a major problem. By characterising the growing environments, it is possible to match crop phenology with the environment and target specific drought environments. This could lead to minimal losses from terminal drought by ensuring the reproductive phase which is most sensitive to drought is reached in a period when soil moisture is not limiting. Short season crops can be grown in areas where drought starts early in the season, whereas longer maturing genotypes can be grown in areas where soil moisture is adequate. Similarly, the framework for developing chickpea ideotype can be used to develop ideotypes of other crops which are important strategies in adapting to adverse environments. 105 8.0: GENERAL DISCUSSION 8.1 Introduction Chickpea is an important legume that provides dietary protein in both human and animal diets. It also ameliorates the soil through atmospheric nitrogen fixation. However, chickpea suffers from terminal drought in many of the areas that it is cultivated. This condition is exacerbated by the fact that climate change may cause an increase in intensity and frequency of droughts in the future. Supplementary irrigation may be used, however 80% of all allocable water is currently already used in agriculture and this option may not always be feasible. Growing chickpea genotypes that have high water use efficiency and can sustain yield under drought environments is a better option. However, the challenge still remains because water use efficiency is a complex trait and not an easy target for plant breeders. This breeding challenge be overcome by identifying secondary traits that are highly heritable and simple to work with as surrogates. A combination of improved genotypes and management options, including tillage practices, can help increase water use efficiency and sustain yields under water limited conditions. This thesis investigated; i) water use, WUE and yield variation in chickpea genotypes, ii) the basis of chickpea yield under water limited field conditions, iii) effect of genotype by environment by management interactions on chickpea phenotypic stability and iv) development of a drought tolerant chickpea ideotype for the Australian grain belt (Figure 8.1). 106 Topic WUE Scope Breeding Aims Test genetic variation; Chapter 4 Outcomes • • • • • Physiological Understand physiological basis Chapter 5 Management Understand G xExM Chapter 6 Modelling Develop ideotype Chapter 7 Identified genotypes with high WUE Identified drought tolerant and susceptible genotypes for use in breeding programs Identified best management practices/growing environments Identified the physiological basis of drought tolerance under water limited conditions Developed chickpea ideotype Figure 8.1: Schematic presentation of the scope, aims and findings of the present study 107 Table 8.1: Thesis summary with objectives, key findings and outcomes Chapter Objective Key findings Outcomes 4 Elucidate differences in No difference in WU in the tested chickpea Better understanding of WU and WUE efficiency results vary, WU and WUE genotypes but WUE was different more research is needed in a WUE in chickpea genotypes Further enquiry multi-factor level, using 5 Discover the effect of No till generally had higher WUE No till may be more beneficial than diverse soils and water tillage and irrigation on WUE efficiency was higher under till due to increased WUE regimes WUE irrigation Identify drought tolerant Sonali, PBA Slasher and ICCV 96853 Drought tolerant and susceptible Single drought indices may and drought susceptible were drought tolerant and Genesis 079 and chickpea types identified not always be the best genotypes Amethyst are susceptible predictors of drought tolerant genotypes. More indices Identify traits associated A total of 21 phenological, morphological, Better understanding of chickpea should be evaluated under with yield formation physiological and yield component traits yield formation under water different drought intensities under water stressed identified stressed conditions Identify field trait NDVI and chlorophyll content identified Relationship between marker traits markers using drought as good marker traits for drought tolerance and drought tolerance identified conditions indices 108 6 Measure extent of G x E There was significant G x E interaction interaction in chickpea G x E in different environments More diverse environments confirmed need to be included to further test the extent of G x E in Identify possible mega- Two mega-environments were identified environments Discriminating ability and chickpea representativeness of environment identified 7 Identify stable and high The ideal genotype was Sonali, followed Ideal genotypes identified yielding genotypes by ICCV 96853 and PBA Slasher Develop chickpea Chickpea ideotype outperformed Chickpea ideotype design and Further research is needed to ideotype commercial cultivars performance evaluated evaluate the performance of the ideotype under drought Characterise chickpea Three major growing environments were Drought patterns in growing areas growing environments identified identified Identify critical growth Reproductive phase is the most critical Critical stage identified stage for drought damage stage in terms of sensitivity to drought and irrigated conditions 109 8.2 Water use and water use efficiency in chickpea Water use was not significantly different among the genotypes which is consistent with data reported by Brown et al. (1989). Chickpea yield was generally higher under no till compared with the tillage treatment. This may be attributed to the higher moisture levels evidenced in the no-till treatment resulting from soil moisture conservation and storage (Felton et al., 1995) and lower soil temperatures and evaporation due to higher plant residues on the soil surface (Hatfield et al., 2001). Genotypes that had high yield potential under water stressed conditions were drought tolerant with high WUE. Since chickpea is mostly grown on stored soil moisture, it is important to make management decisions that ensure moisture is conserved. No-till provides such an avenue for soil moisture conservation under receding moisture conditions and may be a helpful management option for chickpeas. WUE was higher under no-till than under till conditions with Sonali, ICCV 96853 and PBA having high WUE across all the treatments compared with the other test varieties. The most water efficient genotypes can be used as parents in a breeding program to increase WUE or grown directly by growers. The observed genotypic variation for WUE was generally low and there is a need to diversify the genetic base through germplasm introductions or hybridisation in efforts to breed for high WUE. 8.3 Chickpea yield under water limited conditions Selection efforts for genotypes that are high yielding under both well-watered and water stressed conditions should be done carefully in order to obtain the best genotypes. In the present study, the use of drought indices has been shown to be a useful tool for identifying drought tolerant genotypes that have high yield potential under well-watered conditions and that can sustain yield under water limited conditions. Mean relative performance (MRP), relative efficiency index (REI) and stress tolerance index (STI) were the best of the indices used in this chickpea study for identifying drought tolerant genotypes with a high yield potential. These indices were highly correlated with yield under both well-watered and water stressed environments. Several traits (21 in total) were identified as the main contributors (explained 91% of the total variation) to yield variation under water stressed environments. These traits included phenology (days to 50% flowering, days to last flower and flower duration), morphological (leaf characteristics and plant height), physiological (chlorophyll content and NDVI) and yield components (biomass, harvest index and seed weight). Water deficit conditions generally reduced the expression of these traits. 110 Trait association with drought indices can be used to identify select for drought tolerance in the field. NDVI and chlorophyll content were significantly and positively associated with mean relative performance, relative efficiency index and stress tolerance index. This means that NDVI and chlorophyll content can be used in the field to identify genotypes that are drought tolerant. By identifying such genotypes early in the field, the plant breeder can observe other traits during selection that are not necessarily related to yield but may be of importance to the end user. 8.4 Chickpea phenotypic stability Total rainfall and rainfall distribution plays a key role in yield formation under water limited conditions. In the present study, there was less rainfall in 2014 during the vegetative phase than in 2015 but the yields were higher in 2014. This indicated that the vegetative phase may not be the critical stage for yield formation under rainfed conditions. There was a large seasonal effect (year) that caused much of the variation in yield. In order to understand the effect of environment on yield stability, GGE biplots were used. They identified IRN14 and RFN14 as highly discriminating and representative environments. This indicates that genotype evaluation can be done in these two environments and representative information for the other environments under study will still be obtained. Evaluating phenotypic stability alone without yield potential is not sufficient. It is important to identify genotypes that have high yield potential and are stable across environments. In the present study, Sonali, PBA Slasher and ICCV 96853 were identified as relatively stable and possessing high yield potential. 8.5 Chickpea ideotype Phenological, morphological, physiological and yield component traits were identified and used to construct a chickpea ideotype. Ideotype breeding helps the plant breeder focus selection on important traits and introgress them into the desired background. The constructed plant ideotype performed better than the commercially grown cultivars under a range of environments. Environmental characterisation delineated the Australian grain belt into three major clusters based on soil water deficit ratios. These environments varied in terms of drought and the timing of drought, with the high yielding environment having very little moisture deficit during the reproductive phase. The two other environments had different drought intensities towards the end of the growing season and affected the reproductive phase which was identified as the most sensitive to drought. 111 8.6 Summary of discussions There is continued increase in incidence of drought in areas where chickpea is cultivated. This has called for concerted efforts in addressing this problem. An integrated approach was used in the present study whereby genotypes with high water use efficiency were identified in chickpea varieties commonly grown in farmers’ fields and also used as parents in a breeding program. For the genotypes that are commonly grown under farmer field conditions, stability of yield was accessed under varying conditions to ensure sustainability of yield under diverse environmental conditions. Several physiological traits were also identified to help in the breeding program as well as key target traits in developing the chickpea ideotype. 8.7 Conclusions There was genetic variation for WUE but not for water use. Chickpea genotypes that had high yield potential coupled with WUE performed well under both well-watered and water limited conditions. There was a positive correlation between non stressed chickpea yield and stressed chickpea yield, and as such, selection performed under well-watered conditions should lead to high yields under water stressed conditions. Sonali, PBA Slasher and ICCV 96853 were identified as drought tolerant using three drought indices; namely mean relative performance, relative efficiency index and stress tolerance index. NDVI at early and late podding, as well as chlorophyll content at late podding, can be used as markers to select for drought tolerant genotypes in the field. GGE biplot analysis grouped the growing environments into two mega environments with Sonali, ICCV 96853 and PBA Slasher showing a wider adaptation into the environments. Water use was not different among the genotypes tested but hydraulic conductance was significantly different (P<0.05) for whole plant, and root and stem. Water stress reduced expression of most morphological traits. Sonali, which is drought tolerant, had high hydraulic conductance for whole plant, root and stem, and leaves under water stressed conditions enabling it to quench the transpiration stream. The developed chickpea ideotype outperformed the commercial genotypes tested in silico across a wide range of environments. Sonali was closer to the ideotype and had 76% similarity; hence it is a suitable target to introgress the preferred ideotype traits. 112 In general, chickpea productivity in water stressed environments can be increased by selecting genotypes with high yield potential and high WUE. These genotypes should show drought tolerance and be stable across environments. By targeting secondary traits that confer yield under water stressed environments, and using them to construct chickpea ideotypes which can be matched to the growing environment, yield may be increased. 8.7 Further research  More research is needed to understand chickpea WUE in different tillage systems.  Single drought indices may not be reliable in identification of drought tolerance in chickpea. A large combination of indices should be further tested under different drought intensities to identify the best combination for drought tolerant chickpea genotypes. These indices should also be tested to verify consistency of the identified physiological markers for drought tolerance.  The genomic regions for the 21 identified traits that confer yield under water limited environments need to be identified using molecular tools for further testing and introgression.  There was genotype by environment interaction under different soil and tillage environments. There is a need to test this interaction further by incorporating various sites with different soil types and moisture regimes to see if the interaction is repeatable and thus can be exploited by plant breeders. 113 References AGYEMAN, A., PARKES, E. & PEPRAH, B. 2015. AMMI and GGE biplot analyses of root yield performance of cassava genotypes in forest and coastal ecologies. International Journal of Agricultural Policy and Research, 3, 122-132. AHLWAT, I. P. S. & SHIVAKUMAR, B. G. 2003. Cropping systems research in chickpea. In: MASOOD, A., KUMAR, S. & SINGH, N. B. (eds.) Chickpea Research in India. India Institute of Pulse Research. AHMAD, F. & HYMOWITZ, T. 1993. The fine structure of chickpea (Cicer arietinum L.) chromosomes as revealed by pachytene analysis. Theoretical and Applied Genetics, 86, 637-641. AHMAD, R., QADIR, S., AHMAD, N. & SHAH, K. 2003. Yield potential and stability of nine wheat varieties under water stress conditions. International Journal of Agriculture and Biology, 5, 7-9. AHMED, S., NAWATA, E., HOSOKAWA, M., DOMAE, Y. & SAKURATANI, T. 2002. Alterations in photosynthesis and some antioxidant enzymatic activities of mungbean subjected to waterlogging. Plant Science, 163, 117-123. AHMED, S. U. 2011. Effects of soil water deficit on leaf nitrogen, chlorophylls and spad chlorophyll meter reading on growth stages of soybean. Bangladesh Journal Of Botany, 40, 171-175. AKCURA, M. & CERI, S. 2011. Evaluation of drought tolerance indices for selection of Turkish oat (Avena sativa L.) landraces under various environmental conditions. Zemdirbyste Agriculture, 98, 157-166. AKPONIKPÈ, P. B. I., GÉRARD, B., MICHELS, K. & BIELDERS, C. 2010. Use of the APSIM model in long term simulation to support decision making regarding nitrogen management for pearl millet in the Sahel. European Journal of Agronomy, 32, 144154. ALI, M. H. & TALUKDER, M. S. U. 2008. Increasing water productivity in crop production—A synthesis. Agricultural Water Management, 95, 1201-1213. ANDRADE, F. H., CALVIÑO, P., CIRILO, A. & BARBIERI, P. 2002. Yield responses to narrow rows depend on increased radiation interception. Agronomy Journal, 94, 975980. ANGADI, S. V., MCCONKEY, B. G., CUTFORTH, H. W., MILLER, P. R., ULRICH, D., SELLES, F., VOLKMAR, K. M., ENTZ, M. H. & BRANDT, S. A. 2008. Adaptation of alternative pulse and oilseed crops to the semiarid Canadian prairie: seed yield and water use efficiency. Canadian Journal of Plant Science, 88, 425-438. ANNICCHIARICO, P. 2002. Genotype x Environment Interactions - Challenges and Opportunities for Plant Breeding and Cultivar Recommendations. FAO Plant production and protection paper, 174. ANWAR, M. R., MCKENZIE, B. A. & HILL, G. D. 1999. Water use efficiency of chickpea (Cicer arietinum L.) cultivars in Canterbury: effect of irrigation and sowing date. Agronomy N.Z, 29. ANWAR, M. R., MCKENZIE, B. A. & HILL, G. D. 2003. Water-use efficiency and the effect of water deficits on crop growth and yield of Kabuli chickpea (Cicer arietinum L.) in a cool-temperate subhumid climate. The Journal of Agricultural Science, 141, 285-301. ARAUS, J. L. & CAIRNS, J. E. 2014. Field high-throughput phenotyping: the new crop breeding frontier. Trends in Plant Science, 19, 52-61. ARAUS, J. L., SLAFER, G. A., REYNOLDS, M. & ROYO, C. 2002. Plant breeding and drought in C3 cereals: what should we breed for? Annals of Botany, 89, 925-940. 114 ARAUS, J. L. & TAPIA, L. 1987. Photosynthetic gas exchange characteristics of wheat flag leaf blades and sheaths during grain filling. Plant physiology, 85, 667-673. ARCHONTOULIS, S. V., MIGUEZ, F. E. & MOORE, K. J. 2014. Evaluating APSIM Maize, Soil Water, Soil Nitrogen, Manure, and Soil Temperature Modules in the Midwestern United States. Agronomy Journal, 106, 1025-1040. BACELAR, E. L. V. A., MOUTINHO-PEREIRA, J. M., GONÇALVES, B. M. C., BRITO, C. V. Q., GOMES-LARANJO, J., FERREIRA, H. M. F. & CORREIA, C. M. 2012. Water Use Strategies of Plants Under Drought Conditions. In: AROCA, R. (ed.) Plant Responses to Drought Stress: From Morphological to Molecular Features. Berlin, Heidelberg: Springer Berlin Heidelberg. BARRS, H. D. & WEATHERLEY, P. E. 1962. A re-examination of the relative turgidity technique for estimating water deficits on leaves. In: MOINUDDIN & KHANNACHOPRA, R. (eds.) Osmotic adjustment in chickpea in relation to seed yield and yield parameters. Crop Science Society of America: 2004. BAUER, M. E. 1985. Spectral inputs to crop identification and condition assessment. Proceedings of the IEEE, 73, 1071-1085. BEEBE, S. E., RAO, I. M., CAJIAO, C. & GRAJALES, M. 2008. Selection for drought resistance in common bean also improves yield in phosphorus limited and favorable environments. Crop Science, 48, 582-592. BEECH, D. & LEACH, G. 1988. Response of chickpea accessions to row spacing and plant density on a vertisol on the Darling Downs, south-eastern Queensland. 2. Radiation interception and water use. Animal Production Science, 28, 377-383. BEHBOUDIAN, M. H., MA, Q., TURNER, N. C. & PALTA, J. A. 2001. Reactions of chickpea to water stress: yield and seed composition. Journal of the Science of Food and Agriculture, 81, 1288-1291. BERGER, J. D., TURNER, N. C., SIDDIQUE, K. H. M., KNIGHTS, E. J., BRINSMEAD, R. B., MOCK, I., EDMONDSON, C. & KHAN, T. N. 2004. Genotype by environment studies across Australia reveal the importance of phenology for chickpea (Cicer arietinum L.) improvement. Australian Journal Of Agricultural Research, 55, 1071-1084. BERRADA, A. F., SHIVAKUMAR, B. G. & YADURAJU, N. T. 2007. Chickpea in cropping systems. In: YADAV, S. S., REDDEN, R. J., W., C. & SHARMA, B. (eds.) Chickpea Breeding and management. UK: CAB International. BETRAN, F., BECK, D., BÄNZIGER, M. & EDMEADES, G. 2003. Genetic analysis of inbred and hybrid grain yield under stress and nonstress environments in tropical maize. Crop Science, 43, 807-817. BIMBRAW, A. S. 2016. Use of Conservation Technology for the Improvement in Production of Chickpea in Comparison to Wheat. Current Agricultural Research Journal, 4. BLUM, A. 1983. Genetic and physiological relationships in plant breeding for drought resistance. Agricultural Water Management, 7, 195-205. BLUM, A. 1988. Plant breeding for stress environments, Boca Raton, Fla, CRC Press. BLUM, A. 2005. Drought resistance, water-use efficiency, and yield potential - are they compatible, dissonant, or mutually exclusive? Australian Journal of Agricultural Research, 56, 1159-1168. BLUM, A. 2006. Drought adaptation in cereal crops. In: RIBAUT, J. (ed.) Drought adaptation in cereals. New York: Food Products Press. BLUM, A. & EBERCON, A. 1981. Cell Membrane Stability as a Measure of Drought and Heat Tolerance in Wheat1. Crop Science, 21, 43-47. BOCKUS, W. W. & SHROYER, J. P. 1998. The impact of reduced tillage on soilborne plant pathogens. Annual Review of Phytopathology, 36, 485-500. 115 BONHOMME, R. 2000. Beware of comparing RUE values calculated from PAR vs solar radiation or absorbed vs intercepted radiation. Field Crops Research, 68, 247-252. BORRELL, A. K., MULLET, J. E., GEORGE-JAEGGLI, B., VAN OOSTEROM, E. J., HAMMER, G. L., KLEIN, P. E. & JORDAN, D. R. 2014. Drought adaptation of stay-green sorghum is associated with canopy development, leaf anatomy, root growth, and water uptake. Journal of Experimental Botany, 65, 6251-6263. BOUSLAMA, M. & SCHAPAUGH, W. 1984. Stress tolerance in soybeans. I. Evaluation of three screening techniques for heat and drought tolerance. Crop Science, 24, 933-937. BOYER, J. S. 1982. Plant productivity and environment. Science, 218, 443-448. BRAMLEY, H., TURNER, N. C., TURNER, D. W. & TYERMAN, S. D. 2009. Roles of Morphology, Anatomy, and Aquaporins in Determining Contrasting Hydraulic Behavior of Roots. Plant Physiology, 150, 348-364. BROWN, K. D., SORRELLS, M. E. & COFFMAN, W. R. 1983. A Method for Classification and Evaluation of Testing Environments1. Crop Science, 23. BROWN, S., GREGORY, P., COOPER, P. & KEATINGE, J. 1989. Root and shoot growth and water use of chickpea (Cicer arietinum) grown in dryland conditions: effects of sowing date and genotype. The Journal of Agricultural Science, 113, 41-49. BYERLEE, D. & WHITE, R. Agricultural intensification and diversification through food legumes: Technological and policy options. In: KNIGHT, R., ed. Linking research and marketing opportunities for pulses in the 21st century: proceedings of the Third International Food Legumes Research Conference, 2000 Boston. Kluwer Academic Publishers, 31-46. CANCI, H. & TOKER, C. 2009. Evaluation of Yield Criteria for Drought and Heat Resistance in Chickpea (Cicer arietinum L.). Journal Of Agronomy And Crop Science, 195, 47-54. CARBERRY, P. S. 1996. Assessing the opportunity for increased production of grain legumes in the farming system. Final Report to the Grains Research and Development Corporation Project CSC9. . Australia: Grains Research and Development Corporation. CARTELAT, A., AGATI, G., MOYA, I., CEROVIC, Z. G., GOULAS, Y., MEYER, S., LELARGE, C., PRIOUL, J. L., BARBOTTIN, A., JEUFFROY, M. H. & GATE, P. 2005. Optically assessed contents of leaf polyphenolics and chlorophyll as indicators of nitrogen deficiency in wheat (Triticum aestivum L.). Field Crops Research, 91, 3549. CECCARELLI, S. & GRANDO, S. 1991. Selection environment and environmental sensitivity in barley. Euphytica, 57, 157-167. CENKCI, S., CIĞERCI, İ. H., YıLDıZ, M., ÖZAY, C., BOZDAĞ, A. & TERZI, H. 2010. Lead contamination reduces chlorophyll biosynthesis and genomic template stability in Brassica rapa L. Environmental and Experimental Botany, 67, 467-473. CHALK, P. M. 1998. Dynamics of biologically fixed N in legume-cereal rotations: a review. Australian Journal Of Agricultural Research, 49, 303-316. CHAN, K. Y. & MEAD, J. A. 1988. Surface physical properties of a sandy loam soil under different tillage practices. Australian Journal of Soil Research, 26, 549-559. CHAN, K. Y., ROBERTS, W. P. & HEENAN, D. P. 1992. Organic carbon and associated soil properties of a red earth after 10 years of rotation under different stubble and tillage practices. Australian Journal of Soil Research, 30, 71-83. CHAPMAN, S. C., COOPER, M. & HAMMER, G. L. 2002. Using crop simulation to generate genotype by environment interaction effects for sorghum in water-limited environments. Australian Journal of Agricultural Research, 53, 379-389. 116 CHATURVEDI, S., MISHRA, D., VYAS, P. & MISHRA, N. 2009. Breeding for cold tolerance in chickpea. Trends in Biosciences, 2, 1-4. CHAUHAN, Y., WRIGHT, G., RACHAPUTI, N. & MCCOSKER, K. 2008. Identifying chickpea homoclimes using the APSIM chickpea model. Crop and Pasture Science, 59, 260-269. CHAVES, M. M. 1991. Effects of water deficits on carbon assimilation. Journal of Experimental Botany, 42, 1-16. CHENU, K., COOPER, M., HAMMER, G. L., MATHEWS, K. L., DRECCER, M. F. & CHAPMAN, S. C. 2011. Environment characterization as an aid to wheat improvement: Interpreting genotype-environment interactions by modelling waterdeficit patterns in North-Eastern Australia. Journal of Experimental Botany, 62, 17431755. CLARKE, J. M., DEPAUW, R. M. & TOWNLEY-SMITH, T. F. 1992. Evaluation of methods for quantification of drought tolerance in wheat. Crop Science, 32, 723-728. COBOS, M. J., RUBIO, J., FERNANDEZ-ROMERO, M. D., GARZA, R., MORENO, M. T., MILLAN, T. & GIL, J. 2007. Genetic analysis of seed size, yield and days to flowering in a chickpea recombinant inbred line population derived from a Kabuli x Desi cross. Annals of Applied Biology, 151, 33-42. CONDON, A. G., RICHARDS, R. A., REBETZKE, G. J. & FARQUHAR, G. D. 2002. Improving intrinsic water-use efficiency and crop yield. Crop Science, 42, 122-131. CONDON, A. G., RICHARDS, R. A., REBETZKE, G. J. & FARQUHAR, G. D. 2004. Breeding for high water-use efficiency. Journal of Experimental Botany, 55, 24472460. COOK, R. J. & HAGLUND, W. A. 1991. Wheat yield depression associated with conservation tillage caused by root pathogens in the soil not phytotoxins from the straw. Soil Biology and Biochemistry, 23, 1125-1132. DALAL, R. C. 1989. Long-term effects of no-tillage, crop residue, and nitrogen application on properties of a vertisol. Soil Science Society Of America Journal, 53, 1511-1515. DALAL, R. C., STRONG, W. M., DOUGHTON, J. A., WESTON, E. J., MCNAMARA, G. T. & COOPER, J. E. 1997. Sustaining productivity of a Vertisol at Warra, Queensland, with fertilisers, no-tillage or legumes .4. Nitrogen fixation, water use and yield of chickpea. Australian Journal Of Experimental Agriculture, 37, 667-676. DANG, Y. P., SEYMOUR, N. P., WALKER, S. R., BELL, M. J. & FREEBAIRN, D. M. 2015. Strategic tillage in no-till farming systems in Australia’s northern grainsgrowing regions: I. Drivers and implementation. Soil and Tillage Research, 152, 104114. DARVISHZADEH, R., PIRZAD, A., BERNOUSI, I., ABDOLLAHI MANDOULAKANI, B., AZIZI, H., AKHONDI, N., POORMOHAMMAD KIANI, S. & SARRAFI, A. 2011. Genetic properties of drought tolerance indices in sunflower. Acta Agriculturae Scandinavica, Section B-Soil & Plant Science, 61, 593-601. DAVIES, S. L., TURNER, N. C., SIDDIQUE, K. H. M., PLUMMER, J. A. & LEPORT, L. 1999. Seed growth of desi and kabuli chickpea (Cicer arietinum L.) in a short-season Mediterranean-type environment. Australian Journal of Experimental Agriculture, 39, 181-188. DE BOER, R. F., STEED, G. R., KOLLMORGEN, J. F. & MACAULEY, B. J. 1993. Effects of rotation, stubble retention and cultivation on take-all and eyespot of wheat in northeastern Victoria, Australia. Soil & Tillage Research, 25, 263-280. DE WET, J. M. J., ZEVEN, A. C. & ZHUKOVSKIĬ, P. M. 1982. Dictionary of cultivated plants and their regions of diversity, excluding most ornamentals, forest trees and lower plants, Wageningen, Pudoc. 117 DERPSCH, R. 2005. The extent of conservation agriculture adoption worldwide: Implications and impact. III World congess on conservation agriculture: Linking production, livelihoods and conservation. Nairobi. DONALD, C. M. 1968. The breeding of crop ideotypes. Euphytica, 17, 385-403. DORAN, J. W., ELLIOTT, W. T. & PAUSTIAN, K. 1998. Soil microbial activity, nitrogen cycling, and long-term changes in organic carbon pools as related to fallow tillage management. Soil & Tillage Research, 49, 3-18. DOUGLAS RAMSEY, R., FALCONER, A. & JENSEN, J. R. 1995. The relationship between NOAA-AVHRR NDVI and ecoregions in Utah. Remote Sensing of Environment, 53, 188-198. EL-HENDAWY, S. E., HASSAN, W. M., AL-SUHAIBANI, N. A. & SCHMIDHALTER, U. 2017. Spectral assessment of drought tolerance indices and grain yield in advanced spring wheat lines grown under full and limited water irrigation. Agricultural Water Management, 182, 1-12. EL-RAWY, M. A. & HASSAN, M. I. 2014. Effectiveness of drought tolerance indices to identify tolerant genotypes in bread wheat (Triticum aestivum L.). Journal of Crop Science and Biotechnology, 17, 255-266. EXNER, D. N. & CRUSE, R. M. 1993. Interseeded forage legume potential as winter ground cover, nitrogen-source, and competitor. In: BANIK, P., MIDYA, A., SARKAR, B. K. & GHOSE, S. S. (eds.) Wheat intercropping systems in an additive series experiments: Advantages and weed smothering. European Journal of Agronomy. FANG, X. W., TURNER, N. C., LI, F. M. & SIDDIQUE, K. H. M. 2011. An early transient water deficit reduces flower number and pod production but increases seed size in chickpea (Cicer arietinum L.). Crop & Pasture Science, 62, 481-487. FANG, X. W., TURNER, N. C., YAN, G. J., LI, F. M. & SIDDIQUE, K. H. M. 2010. Flower numbers, pod production, pollen viability, and pistil function are reduced and flower and pod abortion increased in chickpea (Cicer arietinum L.) under terminal drought. Journal of Experimental Botany, 61, 335-345. FAO 1983. Guidelines: Land evaluation for rainfed agriculture. FAO Soils bulletin. Rome. FAOSTAT 2012. FAOSTAT. In: KASHIWAGI, J., KRISHNAMURTHY, L., GAUR, P. M., UPADHYAYA, H. D., VARSHNEY, R. K. & TOBITA, S. (eds.) Traits of relevance to improve yield under terminal drought stress in chickpea (C. Arietinum L.). Field Crops Research. FAROOQ, M., WAHID, A., KOBAYASHI, N., FUJITA, D. & BASRA, S. M. A. 2009. Plant drought stress: effects, mechanisms and management. Agronomy For Sustainable Development, 29, 185-212. FARQUHAR, G. D., EHLERINGER, J. R. & HUBICK, K. T. 1989. Carbon Isotope Discrimination and Photosynthesis. Annual Review of Plant Biology, 40, 503-537. FARQUHAR, G. D. & RICHARDS, R. A. 1984. Isotopic Composition of Plant Carbon Correlates With Water-Use Efficiency of Wheat Genotypes. Australian Journal of Plant Physiology, 11, 539. FARSHADFAR, E., MOHAMMADI, R., FARSHADFAR, M. & DABIRI, S. 2013. Relationships and repeatability of drought tolerance indices in wheat-rye disomic addition lines. Australian Journal of Crop Science, 7, 130. FELTON, W. L., MARCELLOS, H., ALSTON, C., MARTIN, R. J., BACKHOUSE, D., BURGESS, L. W. & HERRIDGE, D. F. 1998. Chickpea in wheat-based cropping systems of northern New South Wales - II. Influence on biomass, grain yield, and crown rot in the following wheat crop. Australian Journal Of Agricultural Research, 49, 401-407. 118 FELTON, W. L., MARCELLOS, H. & MARTIN, R. J. 1995. A comparison of three fallow management strategies for the long-term productivity of wheat in northern New South Wales. Australian Journal Of Experimental Agriculture, 35, 915-921. FERNANDEZ, G. C. J. Effective selection criteria for assessing stress tolerance. In: KUO, C. G., ed. Proceedings of the international symposium on adaptation of vegetables and other food crops in temperate and water stress, 1992 Tainan, Taiwan. FISCHER, R. & MAURER, R. 1978. Drought resistance in spring wheat cultivars. I. Grain yield responses. Australian Journal of Agricultural Research, 29, 897-912. FISCHER, R. A. 1981. Optimizing the use of water and nitrogen through breeding of crops. Plant and Soil, 58, 249-278. FOWLER, R. & ROCKSTROM, J. 2001. Conservation tillage for sustainable agriculture. Soil & Tillage Research, 61, 93-108. FOX, T. C. & GEIGER, D. R. 1986. Osmotic Response of Sugar Beet Source Leaves at CO2 Compensation Point. Plant physiology, 80, 239-241. FOYER, C. H., LAM, H.-M., NGUYEN, H. T., SIDDIQUE, K. H., VARSHNEY, R. K., COLMER, T. D., COWLING, W., BRAMLEY, H., MORI, T. A. & HODGSON, J. M. 2016. Neglecting legumes has compromised human health and sustainable food production. Nature Plants, 2, 16112. FOYER, C. H. & NOCTOR, G. 2000. Oxygen processing in photosynthesis: regulation and signalling. New Phytologist, 146, 359-388. FREEBAIRN, D. M., LITTLEBOY, M., SMITH, G. D. & COUGHLAN, K. J. 1991. Optimising soil surface management in response to climatic risk. Climatic risk in crop production. Proc. international symposium, Brisbane, 1990, 283-305. GAN, Y. T., WARKENTIN, T. D., BING, D. J., STEVENSON, F. C. & MCDONALD, C. L. 2010. Chickpea water use efficiency in relation to cropping system, cultivar, soil nitrogen and Rhizobial inoculation in semiarid environments. Agricultural Water Management, 97, 1375-1381. GANGOPADHYAYA, M. & SARKER, R. P. 1965. Influence of rainfall distribution on the yield of wheat crop. Agricultural Meteorology, 2, 331-350. GANJEALI, A., PORSA, H. & BAGHERI, A. 2011. Assessment of Iranian chickpea (Cicer arietinum) germplasm for drought tolerance. Agricultural Water Management, 98, 1477-1484. GARRIDO, R. & LOPEZ-BELLIDO, L. 2001. Effects of crop rotation and nitrogen fertilization on soil nitrate and wheat yield under rainfed Mediterranean conditions. Agronomie, 21, 509-516. GAUCH, H. & ZOBEL, R. W. 1997. Identifying mega-environments and targeting genotypes. Crop Science, 37, 311-326. GAUCH, H. G., KANG, M. S., MAGARI, R. & KANG, M. 1996. New developments in selecting for phenotypic stability in crop breeding. Genotype-by-environment interaction. CRC Press. GAUR, P., KUMAR, J., GOWDA, C., PANDE, S., SIDDIQUE, K., KHAN, T., WARKENTIN, T., CHATURVEDI, S., THAN, A. & KETEMA, D. Breeding chickpea for early phenology: perspectives, progress and prospects. Proceedings of the Fourth International Food Legumes Research Conference, , 2008 New Delhi, India. GAVUZZI, P., RIZZA, F., PALUMBO, M., CAMPANILE, R., RICCIARDI, G. & BORGHI, B. 1997. Evaluation of field and laboratory predictors of drought and heat tolerance in winter cereals. Canadian Journal of Plant Science, 77, 523-531. GEERVANI, P. Utilization of chickpea in India and scope for novel and alternative uses. In: ICRISAT, ed. Uses of tropical grain legumes: proceedings of a Consultants Meeting, 119 27-30 March 1989, 1991 ICRISAT Center, India. Patancheru A.P. 502 324, India. ICRISAT (International Crops Research Institute for the Semi -Ar id Tropics), 47-56. GHOLINEZHAD, E., DARVISHZADEH, R. & BERNOUSI, I. 2014. Evaluation of Drought Tolerance Indices for Selection of Confectionery Sunflower. Notulae Botanicae Horti Agrobotanici Cluj-Napoca, 42, 187. GHOSH, R., SHARMA, M., TELANGRE, R. & PANDE, S. 2013. Occurrence and Distribution of Chickpea Diseases in Central and Southern Parts of India. American Journal of Plant Sciences, 4, 940-944. GILLER, K. E. & CADISCH, G. 1995. Future benefits from biological nitrogen-fixation - an ecological approach to agriculture. In: BANIK, P., MIDYA, A., SARKAR, B. K. & GHOSE, S. S. (eds.) Wheat and chickpea intercropping systems in an additive series experiment: Advantages and weed smothering. European Journal of Agronomy. GODWIN, R. 1990. Agricultural engineering in development: tillage for crop production in areas of low rainfall, Rome, Food and Agriculture Organization. GREN, E. & QUIST, G. 1985. Effects of drought on photosynthesis, chlorophyll fluorescence and photoinhibition susceptibility in intact willow leaves. Planta, 166, 380-388. GUNES, A., INAL, A., ADAK, M. S., BAGCI, E. G., CICEK, N. & ARASLAN, F. 2008. Effect of drought stress implemented at pre- or post-anthesis stage on some physiological parameters as screening criteria in chickpea cultivars. Journal of plant physiology, 55, 59-67. HAIGH, B., ELIAS, N., HARDEN, S., MARTIN, R. & BELFIELD, S. Factors affecting chickpea yields in northern NSW. In: KNIGHTS, T. & RYAN, M., eds. Chickpeas in farming systems Focus 2005 - a National Conference on Chickpeas in Farming Systems, 2005 Goondiwindi, Queensland, Australia. Pulse Australia Publishers. HAMAYOON, R., KHAN, H., NAZ, L., MUNIR, I., ARIF, M., KHALIL, I. A. & KHAN, A. Z. 2011. Performance of chickpea genotypes under two different environmental conditions. African Journal of Biotechnology, 10, 1534. HAMMER, G., COOPER, M., TARDIEU, F., WELCH, S., WALSH, B., VAN EEUWIJK, F., CHAPMAN, S. & PODLICH, D. 2006. Models for navigating biological complexity in breeding improved crop plants. Trends in Plant Science, 11, 587-593. HAMMER, G. L., VAN OOSTEROM, E. J., CHAPMAN, S. C. & MCLEAN, G. The economic theory of water and nitrogen dynamics and mangement in field crops. In: BORRELL, A. K. & HENZEL, R. G., eds. Proceedings of the fourth Australian sorghum conference, 2001 Kooralbyn, Queensland. Range Media Pty Ltd. HAMWIEH, A. & IMTIAZ, M. 2015. Identifying water-responsive and drought-tolerant chickpea genotypes. Crop and Pasture Science, 66, 1003-1011. HANSON, J. D., LIEBIG, M. A., MERRILL, S. D., TANAKA, D. L., KRUPINSKY, J. M. & STOTT, D. E. 2007. Dynamic cropping systems: Increasing adaptability amid an uncertain future. Agronomy Journal, 99, 939-943. HARRIS, D., KHANAL, L., JOSHI, K. D., MUSA, A. M., RAO, J. V. D.-K. K., KANKAL, M. & RASHID, A. 2005. Impacts of replacing rice fallows with productive crops in Bangladesh, Nepal and eastern India. In: BERRADA, A. F., SHIVAKUMAR, B. G. & YADURAJU, N. T. (eds.) Chickpea in cropping systems. UK: CAB International. HATFIELD, J. L., SAUER, T. J. & PRUEGER, J. H. 2001. Managing soils to achieve greater water use efficiency: A review. Agronomy Journal, 93, 271-280. HAY, R. 1995. Harvest index: a review of its use in plant breeding and crop physiology. Annals Of Applied Biology, 126, 197-216. 120 HEENAN, D. P. & TAYLOR, A. C. 1995. Soil-ph decline in relation to rotation, tillage, stubble retention and nitrogen-fertilizer in SE australia. Soil Use and Management, 11, 4-9. HELBAEK, H. 1970. The plant husbandry at Hacilar. In: MELLAART, J. (ed.) Excavation at Hacilar. Edinburgh, UK: Edinburgh University Press. HERRIDGE, D. F., MARCELLOS, H., FELTON, W. L., TURNER, G. L. & PEOPLES, M. B. 1995. Chickpea increases soil-N fertility in cereal systems through nitrate sparing and N2 fixation. Soil Biology and Biochemistry, 27, 545-551. HOBBS, P. R. 2007. Conservation agriculture: what is it and why is it important for future sustainable food production? The Journal of Agricultural Science, 145, 127. HORN, C., BIRCH, C., DALAL, R. & DOUGHTON, J. 1996. Sowing time and tillage practice affect chickpea yield and nitrogen fixation. 1. Dry matter accumulation and grain yield. Australian Journal of Experimental Agriculture, 36, 695-700. HOSSAIN, S. A., STRONG, W. M., WARING, S. A., DALAL, R. C. & WESTON, E. J. 1996. Comparison of legume-based cropping systems at Warra, Queensland. 2. Mineral nitrogen accumulation and availability to the subsequent wheat crop. Australian Journal of Soil Research, 34, 289-297. HSIAO, T. C. 1973. Plant Responses to Water Stress. Annual Review of Plant Physiology, 24, 519-570. HULSE, J. H. Nature, composition and utilization of grain legumes. In: ICRISAT, ed. Uses of tropical Legumes: Proceedings of a Consultants' Meeting, 27-30 March 1989, 1991 ICRISAT Center, Patancheru, A.P. 502 324, India. ICRISAT, 11-27. IMTIAZ, M., MALHOTRA, R. S., SINGH, M. & ARSLAN, S. 2013. Identifying high yielding, stable chickpea genotypes for spring sowing: Specific adaptation to locations and sowing seasons in the Mediterranean region. Crop Science, 53, 14721480. IPCC 2007. Climate change 2007: The Physical Science Basis., Cambridge, United Kingdom and New York, NY, USA, Cambridge University Press. JACKSON, P., ROBERTSON, M., COOPER, M. & HAMMER, G. 1996. The role of physiological understanding in plant breeding; from a breeding perspective. Field Crops Research, 49, 11-37. JACKSON, R. D., REGINATO, R. J. & IDSO, S. B. 1977. Wheat canopy temperature: A practical tool for evaluating water requirements. Water Resources Research, 13, 651. JAIN, J. K. 1975. Breeding for yield and other attributes in grain legumes. Indian Journal of Genetics, 35, 169–187. JAN, A., AMANULLAH, AKBAR, H. & BLASER, B. C. 2012. Chickpea response to tillage system and phosphorus management under dryland conditions. Journal of Plant Nutrition, 35, 64-70. JANGPROMMA, N., SONGSRI, P., THAMMASIRIRAK, S. & JAISIL, P. 2010. Rapid assessment of chlorophyll content in sugarcane using a SPAD chlorophyll meter across different water stress conditions. Asian Journal of Plant Sciences, 9, 368-374. JEYABAL, A. & KUPPUSWAMY, G. 2001. Recycling of organic wastes for the production of vermicompost and its response in rice–legume cropping system and soil fertility. European Journal of Agronomy, 15, 153-170. JOHNSON, M. V. V., KINIRY, J. R. & BURSON, B. L. 2010. Ceptometer Deployment Method Affects Measurement of Fraction of Intercepted Photosynthetically Active Radiation. Agronomy Journal, 102, 1132-1137. JOHNSON, S. S. & GEADELMANN, J. L. 1989. Influence of water-stress on grain-yield response to recurrent selection in maize. Crop Science, 29, 558-564. 121 JOHNSON, V. A. New Horizons in Cereal Production. In: GALLAGHER, E. J. & KELLOGG, W. K., eds. Cereal production: proceedings of the Second International Summer School in Agriculture 1984 Boston;London;. Butterworths in association with the Royal Dublin Society, 1-11. JONES, O. R., HAUSER, V. L. & POPHAM, T. W. 1994. No-tillage effects on infiltration, runoff, and water conservation on dryland. Transactions of the ASAE, 37, 473-479. KANOUNI, H., FARAYEDI, Y., SAEID, A. & SABAGHPOUR, S. H. 2015. Stability Analyses for Seed Yield of Chickpea (Cicer arietinum L.) Genotypes in the Western Cold Zone of Iran. Journal of Agricultural Science, 7, 219. KASHIWAGI, J., KRISHNAMURTHY, L., CROUCH, J. H. & SERRAJ, R. 2006a. Variability of root length density and its contributions to seed yield in chickpea ( Cicer arietinum L.) under terminal drought stress. Field Crops Research, 95, 171-181. KASHIWAGI, J., KRISHNAMURTHY, L., CROUCH, J. H. & SERRAJ, R. 2006b. Variability of root length density and its contributions to seed yield in chickpea (Cicer arietinum L.) under terminal drought stress. Field Crops Research, 95, 171-181. KASHIWAGI, J., KRISHNAMURTHY, L., GAUR, P. M., UPADHYAYA, H. D., VARSHNEY, R. K. & TOBITA, S. 2013. Traits of relevance to improve yield under terminal drought stress in chickpea (C. arietinum L.). Field Crops Research, 145, 8895. KASHIWAGI, J., KRISHNAMURTHY, L., PURUSHOTHAMAN, R., UPADHYAYA, H. D., GAUR, P. M., GOWDA, C. L. L., ITO, O. & VARSHNEY, R. K. 2015. Scope for improvement of yield under drought through the root traits in chickpea (Cicer arietinum L.). Field Crops Research, 170, 47-54. KASHIWAGI, J., KRISHNAMURTHY, L., SINGH, S., GAUR, P. M., UPADHYAYA, H. D., PANWAR, J. D. S., BASU, P. S., ITO, O. & TOBITA, S. 2006c. Relationship between transpiration efficiency and carbon isotope discrimination in chickpea (C. arietinum L.). SAT Agric. Res., 2, 1-3. KASHIWAGI, J., KRISHNAMURTHY, L., UPADHYAYA, H. D. & GAUR, P. M. 2008. Rapid screening technique for canopy temperature status and its relevance to drought tolerance improvement in chickpea. Journal of SAT Agricultural Research, 6, 105-4. KASHIWAGI, J., KRISHNAMURTHY, L., UPADHYAYA, H. D., KRISHNA, H., CHANDRA, S., VADEZ, V. & SERRAJ, R. 2005. Genetic variability of droughtavoidance root traits in the mini-core germplasm collection of chickpea (Cicer arietinum L.). Euphytica, 146, 213-222. KASSAM, A. 1981. Climate, soil and land resources in North Africa and West Asia. Soil Water and Nitrogen in Mediterranean-Type Environments, 1-29. KATO, Y., KAMOSHITA, A. & YAMAGISHI, J. 2008. Preflowering Abortion Reduces Spikelet Number in Upland Rice (Oryza sativa L.) under Water Stress Crop Science, 48, 2389-2395. KEATING, B. A., CARBERRY, P. S., HAMMER, G. L., PROBERT, M. E., ROBERTSON, M. J., HOLZWORTH, D., HUTH, N. I., HARGREAVES, J. N. G., MEINKE, H., HOCHMAN, Z., MCLEAN, G., VERBURG, K., SNOW, V., DIMES, J. P., SILBURN, M., WANG, E., BROWN, S., BRISTOW, K. L., ASSENG, S., CHAPMAN, S., MCCOWN, R. L., FREEBAIRN, D. M. & SMITH, C. J. 2003. An overview of APSIM, a model designed for farming systems simulation. European Journal of Agronomy, 18, 267-288. KHALILI, M., NAGHAVI, M. R., ABOUGHADAREH, A. P. & TALEBZADEH, S. J. 2012. Evaluating of drought stress tolerance based on selection indices in spring canola cultivars (Brassica napus L.). Journal of Agricultural Science, 4, 78. 122 KHANNA-CHOPRA, R. & SINHA, S. What limits the yield of pulses? Plant processes or plant type. Proceedings of the international congress in plant physiology, 1990. 268278. KHANNA-CHOPRA, R. & SINHA, S. K. 1987. Chickpea: Physiological aspects of growth and yield. In: SAXENA, M. C. & SINGH, K. B. (eds.) The Chickpea. Wallingford, UK: CAB International. KHOLOVÁ, J., MCLEAN, G., VADEZ, V., CRAUFURD, P. & HAMMER, G. L. 2013. Drought stress characterization of post-rainy season (rabi) sorghum in India. Field Crops Research, 141, 38-46. KHUSH, G. S. 1995. Breaking the yield frontier of rice. GeoJournal, 35, 329-332. KINIRY, J. R., SIMPSON, C. E., SCHUBERT, A. M. & REED, J. D. 2005. Peanut leaf area index, light interception, radiation use efficiency, and harvest index at three sites in Texas. Field Crops Research, 91, 297-306. KIRSCHENMANN, L. F. 2007. Potential for a new generation of biodiversity in agroecosystems of the future. Agronomy Journal, 99, 373-376. KNAPP, S. J. & BRIDGES, W. C., JR. 1987. Confidence interval estimators for heritability for several mating and experiment designs. Theoretical and Applied Genetics, 73, 759-63. KOROBOV, R. M. & RAILYAN, V. Y. 1993. Canonical correlation relationships among spectral and phytometric variables for 20 winter-wheat fields. Remote Sensing of Environment, 43, 1-10. KRISHNAMURTHY, L., JOHANSEN, C. & SETHI, S. C. 1999. Investigation of factors determining genotypic differences in seed yield of non-irrigated and irrigated chickpeas using a physiological model of yield determination. Journal of Agronomy and Crop Science, 183, 9-17. KRISHNAMURTHY, L., KASHIWAGI, J., GAUR, P. M., UPADHYAYA, H. D. & VADEZ, V. 2010. Sources of tolerance to terminal drought in the chickpea (Cicer arietinum L.) minicore germplasm. Field Crops Research, 119, 322-330. KRISHNAMURTHY, L., SERRAJ, R., KASHIWAGI, J., PANWAR, J. D. S., KOTESWARA RAO, Y. & KUMAR, J. 2004. Multilocation analysis of yield and tield components of chickpea (Cicer arietinum L.) mapping population under terminal drought. Indian journal of Pulses Research, 17, 17-24. KRISHNAMURTHY, L., VADEZ, V., VARSHNEY, R. K., KASHIWAGI, J., TOBITA, S., ITO, O., UPADHYAYA, H. D., GOWDA, C. L. L., GAUR, P. M., SHESHSHAYEE, M. S. & SINGH, S. 2013. Variation in carbon isotope discrimination and its relationship with harvest index in the reference collection of chickpea germplasm. Functional Plant Biology 40, 1350-1361. KROUMA, A. 2010. Plant water relations and photosynthetic activity in three Tunisian chickpea (Cicer arietinum L.) genotypes subjected to drought. Turkish Journal Of Agriculture And Forestry, 34, 257-264. KRUPINSKY, J. M., BAILEY, K. L., MCMULLEN, M. P., GOSSEN, B. D. & TURKINGTON, T. K. 2002. Managing plant disease risk in diversified cropping systems. Agronomy journal, 94, 198-209. KULIK, M. 1962. Agroclimatic indices of drought. Compendium of abridged reports to the Second Session of CagM (WMO). Hydrometeorological Publishing. Moscú, 75-81. KUMAR, S., DWIVEDI, S., SINGH, S., JHA, S., LEKSHMY, S., ELANCHEZHIAN, R., SINGH, O. & BHATT, B. 2014. Identification of drought tolerant rice genotypes by analysing drought tolerance indices and morpho-physiological traits. SABRAO Journal of Breeding & Genetics, 46. 123 LAKE, L., CHENU, K. & SADRAS, V. O. 2016. Patterns of water stress and temperature for Australian chickpea production. Crop and Pasture Science, 67, 204-215. LAL, R. 1999. Long-term tillage and wheel traffic effects on soil quality for two central Ohio soils. Journal Of Sustainable Agriculture, 14, 67-84. LAN, J. 1998. Comparison of evaluating methods for agronomic drought resistance in crops. Acta Agric Boreali-occidentalis Sinica, 7, 85-87. LAURILA, H., MÄKELÄ, P., KLEEMOLA, J. & PELTONEN, J. 2012. A comparative ideotype, yield component and cultivation value analysis for spring wheat adaptation in Finland. Agricultural and Food Science, 21, 25. LAUTER, D. J. & MUNNS, D. N. 1986. Salt resistance of chickpea genotypes in solutions salinized with NaCl or Na2SO4. Plant and Soil, 95, 271-279. LEPORT, L., TURNER, N. C., DAVIES, S. L. & SIDDIQUE, K. H. M. 2006. Variation in pod production and abortion among chickpea cultivars under terminal drought. European Journal of Agronomy, 24, 236-246. LEPORT, L., TURNER, N. C., FRENCH, R. J., BARR, M. D., DUDA, R., DAVIES, S. L., TENNANT, D. & SIDDIQUE, K. H. M. 1999. Physiological responses of chickpea genotypes to terminal drought in a Mediterranean-type environment. European Journal of Agronomy, 11, 279-291. LEPORT, L., TURNER, N. C., FRENCH, R. J., TENNANT, D., THOMSON, B. D. & SIDDIQUE, K. H. M. 1998. Water relations, gas exchange and growth of cool-season grain legumes in a Mediterranean-type environment. European Journal of Agronomy, 9, 295-303. LI, H. W., GAO, H. W., WU, H. D., LI, W. Y., WANG, X. Y. & HE, J. 2007. Effects of 15 years of conservation tillage on soil structure and productivity of wheat cultivation in northern China. Australian Journal of Soil Research, 45, 344-350. LI, R.-H., GUO, P.-G., MICHAEL, B., STEFANIA, G. & SALVATORE, C. 2006. Evaluation of Chlorophyll Content and Fluorescence Parameters as Indicators of Drought Tolerance in Barley. Agricultural Sciences in China, 5, 751-757. LIN, C.-S., BINNS, M. R. & LEFKOVITCH, L. P. 1986. Stability analysis: where do we stand? Crop Science, 26, 894-900. LINN, D. M. & DORAN, J. W. 1984. Effect of Water-Filled Pore Space on Carbon Dioxide and Nitrous Oxide Production in Tilled and Nontilled Soils. Soil Science Society of America 48, 1267-1272. LIU, Y. B., ZHANG, T. G. & WANG, J. 2008. Photosynthesis and metabolite levels in dehydrating leaves of reaumuria soongorica. Acta Biologica Cracoviensia Series Botanica, 50, 19-26. LÓPEZ-BELLIDO, L., FUENTES, M., CASTILLO, J. E. & LÓPEZ-GARRIDO, F. J. 1998. Effects of tillage, crop rotation and nitrogen fertilization on wheat-grain quality grown under rainfed Mediterranean conditions. Field Crops Research, 57, 265-276. LUDLOW, M. M. & MUCHOW, R. C. 1990. A Critical Evaluation of Traits for Improving Crop Yields in Water-Limited Environments. Advances in Agronomy, 43, 107-153. LUPWAYI, N. Z., RICE, W. A. & CLAYTON, G. W. 1998. Soil microbial diversity and community structure under wheat as influenced by tillage and crop rotation. Soil Biology and Biochemistry, 30, 1733-1741. LYNCH, J. 1995. Root Architecture and Plant Productivity. Plant Physiology, 109, 7-13. MAALOUF, F., MUHAMMAD, I., KUMAR, S. & MALHOTRA, R. 2011. Breeding food legumes for enhanced drought and heat tolerance to cope with climate change. Food Security and Climate Change in Dry Areas, 244. 124 MAFAKHERI, A., SIOSEMARDEH, A., BAHRAMNEJAD, B., STRUIK, P. C. & SOHRABI, E. 2010. Effect of drought stress on yield, proline and chlorophyll contents in three chickpea cultivars. Australian journal of Crop Science, 4, 580-585. MALHOTRA, R. S. & SINGH, K. B. 1991. Classification of chickpea growing environments to control genotype by environment interaction. Euphytica, 58, 5-12. MALLIKARJUNA, N., COYNE, C., CHO, S., RYNEARSON, S., RAJESH, P. N., JADHAV, D. R. & MUEHLBAUER, F. J. 2011. Cicer. In: KOLE, C. (ed.) Wild Crop Relatives: Genomic and Breeding resources, Legumes and Forages. Berlin: SpringerVerlag. MANIVANNAN, P., JALEEL, C. A., KISHOREKUMAR, A., SANKAR, B., SOMASUNDARAM, R., SRIDHARAN, R. & PANNEERSELVAM, R. 2007. Changes in antioxidant metabolism of Vigna unguiculata (L.) Walp. by propiconazole under water deficit stress. Colloids and Surfaces B: Biointerfaces, 57, 69-74. MAQBOOL, M. A., ASLAM, M., ALI, H., SHAH, T. M. & ATTA, B. M. 2015. GGE biplot analysis based selection of superior chickpea (Cicer arietinum L.) inbred lines under variable water environments. Pakistan Journal of Botany, 47, 1901-1908. MARIANO, C. C., PIETRAGALLA, J. & REYNOLDS, M. 2012. Canopy temperature and plant water relations traits. In: REYNOLDS, M., PASK, A. & MULLAN, D. (eds.) Physiological breeding I: Interdisplinary approaches to improve crop adaptation. Mexico: CIMMYT. MARINHO, C. D., GRAVINA, G. D. A., SEBASTIÃO, L. C. D. A., ALMEIDA, N. C. D., DAHER, R. F., BRASILEIRO, B. P., PAULA, T. O. M. D. & AMARAL JÚNIOR, A. T. D. 2014. Indexes in the comparison of pre-commercial genotypes of common bean. Ciência Rural, 44, 1159-1165. MARKWELL, J., OSTERMAN, J. C. & MITCHELL, J. L. 1995. Calibration of the Minolta SPAD-502 leaf chlorophyll meter. Photosynthesis research, 46, 467-472. MARLEY, J. M. & LITTLER, J. W. 1989. Winter Cereal Production On The Darling Downs - An 11 Year Study Of Fallowing Practices. Australian Journal Of Experimental Agriculture, 29, 807-827. MARTRE, P., QUILOT-TURION, B., LUQUET, D., MEMMAH, M.-M. O.-S., CHENU, K. & DEBAEKE, P. 2015. Model-assisted phenotyping and ideotype design In: CALDERINI, D. F. (ed.) Crop Physiology (Second Edition). San Diego: Academic Press. MATOS, M. C., CAMPOS, P. S., PASSARINHO, J. A., SEMEDO, J. N., MARQUES, N. M., RAMALHO, J. C. & RICARDO, C. P. 2010. Drought effect on photosynthetic activity, osmolyte accumulation and membrane integrity of two Cicer arietinum genotypes. Photosynthetica, 48, 303-312. MCGARRY, D., BRIDGE, B. J. & RADFORD, B. J. 2000. Contrasting soil physical properties after zero and traditional tillage of an alluvial soil in the semi-arid subtropics. Soil & Tillage Research, 53, 105-115. MCKENZIE, R. H., MIDDLETON, A. B. & BREMER, E. 2006. Fertilizer and rhizobial inoculant responses of chickpea on fallow and stubble sites in southern Alberta. Canadian Journal of Plant Science, 86, 685-692. MERRILL, S. D., TANAKA, D. L., KRUPINSKY, J. M., LIEBIG, M. A. & HANSON, J. D. 2007. Soil Water Depletion and Recharge under Ten Crop Species and Applications to the Principles of Dynamic Cropping Systems. Agronomy Journal, 99, 931. MIELKE, L. N., DORAN, J. W. & RICHARDS, K. A. 1986. Physical environment near the surface of plowed and no-tilled soils. Soil & Tillage Research, 7, 355-366. MISHRA, A. K. & SINGH, V. P. 2010. A review of drought concepts. Journal of Hydrology, 391, 202-216. 125 MITRA, J. 2001. Genetics and genetic improvement of drought resistance in crop plants. Current Science, 80, 758-763. MOCK, J. J. & PEARCE, R. B. 1975. An ideotype of maize. Euphytica, 24, 613-623. MOELLER, C. & REBETZKE, G. 2017. Performance of spring wheat lines near-isogenic for the reduced-tillering ‘tin’ trait across a wide range of water-stress environment-types. Field Crops Research, 200, 98-113. MOHANTY, M., PROBERT, M. E., REDDY, K. S., DALAL, R. C., MISHRA, A. K., SUBBA RAO, A., SINGH, M. & MENZIES, N. W. 2012. Simulating soybean–wheat cropping system: APSIM model parameterization and validation. Agriculture, Ecosystems & Environment, 152, 68-78. MONNEVEUX, P. & RIBAUT, J.-M. 2006. Secondary traits for drought tolerance improvement in cereals. In: RIBAUT, J.-M. (ed.) Drought adaptation in cereals. New York: Food Products Press. MONTES, J. M., MELCHINGER, A. E. & REIF, J. C. 2007. Novel throughput phenotyping platforms in plant genetic studies. Trends in Plant Science, 12, 433-436. MOORE, K. J. 1984. Increased Take-all of Wheat with Direct Drilling in the Pacific Northwest. Phytopathology, 74, 1044. MORAN, M. S., INOUE, Y. & BARNES, E. M. 1997. Opportunities and limitations for image-based remote sensing in precision crop management. Remote Sensing of Environment, 61, 319-346. MORISON, J. I. L., BAKER, N. R., MULLINEAUX, P. M. & DAVIES, W. J. 2008. Improving water use in crop production. Philosophical transactions of the Royal Society of London. Series B, Biological sciences, 363, 639-658. MUNNS, R. & CRAMER, G. R. 1996. Is coordination of leaf and root growth mediated by abscisic acid? Opinion. Plant and Soil, 185, 33-49. MUÑOZ-ROMERO, V., LÓPEZ-BELLIDO, R. J. & LÓPEZ-BELLIDO, L. 2012. The effects of the tillage system on chickpea root growth. Field Crops Research, 128, 7681. NAYYAR, H., KAUR, S., SINGH, S. & UPADHYAYA, H. D. 2006. Differential sensitivity of Desi (small-seeded) and Kabuli (large-seeded) chickpea genotypes to water stress during seed filling: effects on accumulation of seed reserves and yield. Journal of the Science of Food and Agriculture, 86, 2076-2082. NAZARI, L. & PAKNIYAT, H. 2010. Assessment of drought tolerance in barley genotypes. Journal of Applied Science, 10, 151-156. NONAMI, H., WU, Y. & BOYER, J. S. 1997. Decreased Growth-Induced Water Potential (A Primary Cause of Growth Inhibition at Low Water Potentials). Plant Physiology, 114, 501-509. OHASHI, Y., SANEOKA, H. & FUJITA, K. 2000. Effect of water stress on growth, photosynthesis, and photoassimilate translocation in soybean and tropical pasture legume siratro. Soil science and Plant Nutrition, 46, 417-425. PAGE, K., DANG, Y. & DALAL, R. 2013. Impacts of conservation tillage on soil quality, including soil-borne crop diseases, with a focus on semi-arid grain cropping systems. Australasian Plant Pathology, 42, 363-377. PANDE, S., SHARMA, M., GAUR, P. M., BASANDRAI, A. K., KAUR, L., HOODA, K. S., BASANDRAI, D., KIRAN BABU, T., JAIN, S. K. & RATHORE, A. 2013. Biplot analysis of genotype × environment interactions and identification of stable sources of resistance to Ascochyta blight in chickpea (Cicer arietinum L.). Australasian Plant Pathology, 42, 561-571. PARTHASARATHY RAO, P., BIRTHAL, P. S., BHAGAVATULA, S. & BANTILAN, M. C. S. 2010. Chickpea and Pigeonpea Economies in Asia: Facts, Trends and Outlook, 126 Patancheru 502 324, Andhra Pradesh, India, International Crops Research Institute for the Semi-Arid Tropics. PASSIOURA, J. 2006. Increasing crop productivity when water is scarce—from breeding to field management. Agricultural Water Management, 80, 176-196. PASSIOURA, J. B. 1977. Grain yield, harvest index, and water use of wheat. Journal of the Australian Institute of Agricultural Science, 43, 117-120. PASTENES, C., PIMENTEL, P. & LILLO, J. 2005. Leaf movements and photoinhibition in relation to water stress in field-grown beans. Journal of Experimental Botany, 56, 425-433. PATTERSON, H. D. & THOMPSON, R. 1971. Recovery of Inter-Block Information when Block Sizes are Unequal. Biometrika, 58, 545-554. PENG, S., KHUSH, G. S. & CASSMAN, K. G. Evolution of the new plant ideotype for increased yield potential In: CASSMAN, K. G., ed. Breaking the yield barrier: Rice Yield Potential in Favorable Environments, 1994 IRRI. IRRI. PEÑUELAS, J. & FILELLA, I. 1998. Visible and near-infrared reflectance techniques for diagnosing plant physiological status. Trends in Plant Science, 3, 151-156. PIETRAGALLA, J. 2012. Canopy temperature. In: PASK, A. J. D., PIETRAGALLA, J., MULLAN, D. M. & REYNOLDS, M. P. (eds.) Physiological Breeding II: A Field Guide to Wheat Phenotyping. Mexico: CIMMYT. POORMOHAMMAD, K. S., GRIEU, P., MAURY, P., HEWEZI, T., GENTZBITTEL, L. & SARRAFI, A. 2007. Genetic variability for physiological traits under drought conditions and different expression of water stress-associated genes in sunflower (Helianthus annus L.). Theoretical and Applied Genetics, 114, 193-207. PORCH, T. G. & JAHN, M. 2001. Effects of high‐temperature stress on microsporogenesis in heat‐sensitive and heat‐tolerant genotypes of Phaseolus vulgaris. Plant, Cell & Environment, 24, 723-731. PURUSHUKOTHAMAN, R., THUDI, M., KRISHNAMURTHY, L., UPADHYAYA, H. D., KASHIWAGI, J., GOWDA, C. L. L. & VARSHNEY, R. K. 2015. Association of mid-reproductive stage canopy temperature depression with the molecular markers and grain yields of chickpea (Cicer arietinum L.) germplasm under terminal drought. Field Crops Research, 174, 1-11. RADFORD, B. J., KEY, A. J., ROBERTSON, L. N. & THOMAS, G. A. 1995. Conservation tillage increases soil-water storage, soil animal populations, grain-yield, and response to fertilizer in the semiarid subtropics. Australian Journal Of Experimental Agriculture, 35, 223-232. RAHBARIAN, R., KHAVARI-NEJAD, R., GANJEALI, A., BAGHERI, A. & NAJAFI, F. 2011. Drought stress effects on photosynthesis, chlorophyll fluorescence and water relations in tolerant and susceptible chickpea (Cicer arietinum L.) genotypes. Acta Biologica Cracoviensia Series Botanica, 53, 47-56. RAMAMOORTHY, P., LAKSHMANAN, K., UPADHYAYA, H. D., VADEZ, V. & VARSHNEY, R. K. 2016. Shoot traits and their relevance in terminal drought tolerance of chickpea (Cicer arietinum L.). Field Crops Research, 197, 10-27. RANDHAWA, N., KAUR, J., SINGH, S. & SINGH, I. 2014. Growth and yield in chickpea (Cicer arietinum L.) genotypes in response to water stress. African Journal of Agricultural Research, 9, 982-992. RASMUSSON, D. C. 1987. An Evaluation of Ideotype Breeding. Crop Science, 27, 11401146. 127 RASMUSSON, D. C. 1991. A plant breeder's experience with ideotype breeding. Field Crops Research, 26, 191-200. RATHORE, A. L., PAL, A. R. & SAHU, K. K. 1998. Tillage and mulching effects on water use, root growth and yield of rainfed mustard and chickpea grown after lowland rice. Journal of Science, Food and Agriculture, 78, 149-161. REBETZKE, G. J., CONDON, A. G., RICHARDS, R. A. & FARQUHAR, G. D. 2002. Selection for reduced carbon isotope discrimination increases aerial biomass and grain yield of rainfed bread wheat. Crop Science, 42, 739-745. REDDY, B. V., RAMESH, S., REDDY, P. S. & KUMAR, A. A. 2009. 3 Genetic Enhancement for Drought Tolerance in Sorghum. Plant Breeding Reviews, 31, 189. REYAZUL, R. M., ZAMAN-ALLAH, M., SREENIVASULU, N., TRETHOWAN, R. & VARSHNEY, R. K. 2012. Integrated genomics, physiology and breeding approaches for improving drought tolerance in crops. Theoretical Applied Genetics, 125, 625-645. REYNOLDS, M., PASK, A. & MULLAN, D. 2012. Physiological breeding I: interdisciplinary approaches to improve crop adaptation, CIMMYT. REYNOLDS, M., RAJARAM, S. & SAYRE, K. 1999. Physiological and genetic changes of irrigated wheat in the post green revolution period and approaches for meeting projected global demand. Crop Science, 39, 1611-1621. REYNOLDS, M., TRETHOWAN, R., VAN GINKEL, M. & RAJARAN, S. 2001. Application of physiology in wheat breeding. In: REYNOLDS, M., ORTIZMONASTERIO, J. I. & MCNAB, A. (eds.) Application of physiology in wheat breeding. Mexico: CIMMYT. REYNOLDS, M. & TUBEROSA, R. 2008. Translational research impacting on crop productivity in drought-prone environments. Current Opinion in Plant Biology, 11, 171-179. RICHARDS, R. 1996. Defining selection criteria to improve yield under drought. Plant Growth Regulation, 20, 157-166. ROBERTSON, M. J., CARBERRY, P. S., HUTH, N. I., TURPIN, J. E., PROBERT, M. E., POULTON, P. L., BELL, M., WRIGHT, G. C., YEATES, S. J. & BRINSMEAD, R. B. 2002. Simulation of growth and development of diverse legume species in APSIM. Australian Journal of Agricultural Research, 53, 429-446. ROPER, M. M. & GUPTA, V. 1995. Management-practices and soil biota. Australian Journal of Soil Research, 33, 321-339. ROSIELLE, A. A. & HAMBLIN, J. 1981. Theoretical Aspects of Selection for Yield in Stress and Non-Stress Environment1. Crop Science, 21, 943-946. ROTTER, R. P., TAO, F., HOHN, J. G. & PALOSUO, T. 2015. Use of crop simulation modelling to aid ideotype design of future cereal cultivars. Journal of Experimental Botany, 66, 3463-3476. SAHAR, B., AHMED, B., NASERELHAQ, N., MOHAMMED, J. & HASSAN, O. 2016. Efficiency of selection indices in screening bread wheat lines combining drought tolerance and high yield potential. Journal of Plant Breeding and Crop Science, 8, 7286. SARAF, C. S., RUPELA, O. P., HEGDE, D. M., YADAV, R. L., SHIVKUMAR, B. G., BHATTARAI, S., RAZZAQUE, M. A. & SATTAR, M. A. 1998. Biological nitrogen fixation and residual effects of winter grain legumes in rice and wheat cropping systems of the Indo-Gangetic plain. In: KUMAR, J. V. D. K., JOHANSEN, C. & REGO, T. J. (eds.) Residual effects of legumes in rice and wheat cropping systems of the Indo-Gangetic plain. New Delhi, India: Oxford and IBH Publishing. 128 SARMA, B. K., HEMANTARANJAN, A., PATEL, P. K. & SINGH, R. 2011. Growth and antioxidant system under drought stress in Chickpea (Cicer arietinum L.) as sustained by salicylic acid. Journal of Stress Physiology & Biochemistry, 7, 130-144. SARMAH, B. K., ACHERJEE, S. & SHARMA, H. C. 2012. Chickpea: Crop Improvement under Changing Environment Conditions. In: N., T., GILL, S. S. & TUTEJA, R. (eds.) Improving Crop Productivity in Sustainable Agriculture. Wiley-VCH Verlag GmbH & Co. KGaA. SAXENA, M. C., MALHOTRA, R. S. & SINGH, K. B. 1990. Iron-deficiency in chickpea in the Mediterranean region and its control through resistant genotypes and nutrient application. In: EL-BASSAM, N., DAMBROTH, M. & LOUGHMAN, B. C. (eds.) Genetic Aspects of Mineral Nutrition. Dordrecht, The Netherlands: Kluwer Academic. SAXENA, N. P. 2003. Management of drought in chickpea: a holistic approach. In: SAXENA, N. P. (ed.) Management of Agricultural Drought: Agronomic and Genetic Options. Oxford and IBH Publishing, New Delhi. SAXENA, N. P. & JOHANSEN, C. Chickpea ideotypes for genetic enhancement of yield and yield stability in South Asia. In: VAN RHEENEN, H. A. & SAXENA, M. C., eds. Chickpea in the Nineties: Proceedings of the 2nd International Workshop on Chickpea Improvement, 1990 ICRISAT, Hyderabad, India. ICRISAT, 81-85. SAYAR, R., KEMIRA, H., KAMELI, A. & MOSBAHI, M. 2008. Physiological tests as predictive appreciation for drought tolerance in durum wheat. Agronomy Research, 6, 79-90. SAYED, O. H. 2003. Chlorophyll Fluorescence as a Tool in Cereal Crop Research. Photosynthetica, 41, 321-330. SAYRE, K. & GOVAERTS, B. 2012. The principles of conservation agriculture. In: REYNOLDS, M. P., PASK, A. J. D. & MULLAN, D. M. (eds.) Physiological breeding I: Interdisplinary approaches to improve crop adaptation. Mexico: CIMMYT. SCHULZE, E. D. 1986. Carbon Dioxide and Water Vapor Exchange in Response to Drought in the Atmosphere and in the Soil. Annual Review of Plant Physiology, 37, 247-274. SCHUPPLER, U., HE, P.-H., JOHN, P. C. & MUNNS, R. 1998a. Effect of water stress on cell division and Cdc2-like cell cycle kinase activity in wheat leaves. Plant Physiology, 117, 667-678. SCHUPPLER, U., HE, P.-H., JOHN, P. C. L. & MUNNS, R. 1998b. Effect of Water Stress on Cell Division and Cdc2-Like Cell Cycle Kinase Activity in Wheat Leaves1. Plant Physiology, 117, 667-678. SEDGLEY, R. H., SIDDIQUE, K. H. M. & WALTON, G. H. 1990. Chickpea ideotypes for Mediterranean environments. Patancheru: ICRISAT. SEKHON, H. S., SINGH, G., SHARMA, P. & BAINS, T. S. 2010. Water use efficiency unde stress environments. In: YADAV, S. S., MCNEIL, D. L. & REDDEN, R. (eds.) Climate change and management of cool season grain legume crops. Dordrecht, London: Springer. SEMENOV, M. A. & STRATONOVITCH, P. 2013. Designing high-yielding wheat ideotypes for a changing climate. Food and Energy Security, 2, 185-196. SERRAJ, R. 2004. Symbiotic nitrogen fixation : prospects for enhanced application in tropical agriculture, New Delhi, Oxford & IBH Publishing. SHANAHAN, J. F., SCHEPERS, J. S., FRANCIS, D. D., VARVEL, G. E., WILHELM, W. W., TRINGE, J. M., SCHLEMMER, M. R. & MAJOR, D. J. 2001. Use of remotesensing imagery to estimate corn grain yield. Agronomy Journal, 93, 583-589. 129 SIDDIQUE, K. H. M., REGAN, K. L., TENNANT, D. & THOMSON, B. D. 2001. Water use and water use efficiency of cool season grain legumes in low rainfall Mediterranean-type environments. European Journal of Agronomy, 15, 267-280. SIDDIQUE, K. H. M. & SEDGLEY, R. H. 1985. The effect of reduced branching on yield and water use of chickpea (Cicer arietinum L.) in a Mediterranean type environment. Field Crops Research, 12, 251-269. SIDDIQUE, M. R. B., HAMID, A. & ISLAM, M. S. 2000. Drought stress effects on water relations of wheat. Botanical Bulletin of Academia Sinica, 41, 35-39. SILIM, S. & SAXENA, M. 1993. Adaptation of spring-sown chickpea to the Mediterranean basin. I. Response to moisture supply. Field Crops Research, 34, 121-136. SINCLAIR, T. R. 2011. Challenges in breeding for yield for drought. Trends in Plant Science, 16, 289-293. SINGH, B., RAO, K. & SHARMA, H. 2011. Comparison of selection indices to identify sorghum genotypes resistant to the spotted stemborer Chilo partellus (Lepidoptera: Noctuidae). International Journal of Tropical Insect Science, 31, 38-51. SINGH, D. P., PETERS, D. B., SINGH, P. & SINGH, M. 1987. Diurnal patterns of canopy photosynthesis, evapotranspiration and water use efficiency in chickpea (Cicer arietinum L.) under field conditions. Photosynthesis Research, 11, 61-69. SINGH, G. & BHUSHAN, L. S. 1980. Water use, water-use efficiency and yield of dryland chickpea as influenced by P fertilization, stored soil water and crop season rainfall. Agricultural Water Management, 2, 299-305. SINGH, K. B., MALHOTRA, R. S., HALILA, M. H., KNIGHTS, E. J. & VERMA, M. M. 1993. Current status and future strategy in breeding chickpea for resistance to biotic and abiotic stresses. Euphytica, 73, 137-149. SINGH, P. & SRI RAMA, Y. V. 1989. Influence of water deficit on transpiration and radiation use efficiency of chickpea (Cicer arietinum L.). Agricultural and Forest Meteorology, 48, 317-330. SINGH, R. K. & CHAUDHARY, B. 1979. Biometrical methods in quantitative genetic analysis. Biometrical Methods in Quantitative Genetic Analysis. SMITHSON, J. B., THOMPSON, J. A. & SUMMERFIELD, R. J. 1985. Chickpea (Cicer arietinum L.). In: SUMMERFIELD, R. J. & ROBERTS, E. H. (eds.) Grain Legume Crops. London, UK: Collins. SNEDECOR, G. W. & COCHRAN, W. G. 1987. Statistical Methods, Iowa, U.S.A, The Iowa State University Press. SOJKA, R., STOLZY, L. & FISCHER, R. 1981. Seasonal drought response of selected wheat cultivars. Agronomy Journal, 73, 838-845. STAGGENBORG, S. A. & VANDERLIP, R. L. 2005. Crop simulation models can be used as dryland cropping systems research tools. Agronomy Journal, 97, 378-384. SUMMERFIELD, R., HADLEY, P., ROBERTS, E., MINCHIN, F. & RAWSTHORNE, S. 1984. Sensitivity of Chickpeas (Cicer arietinum L.) to Hot Temperatures during the Reporductive Period. Experimental Agriculture, 20, 77-93. TALEBI, R., FAYAZ, F. & NAJI, A. 2009. Effective selection criteria for assessing drought stress tolerance in durum wheat (Triticum durum Desf.). General And Applied Plant Physiology, 35, 64-74. TARDIEU, F. 2003. Virtual plants: modelling as a tool for the genomics of tolerance to water deficit. Trends in Plant Science, 8, 9-14. TEILLET, P. M. 1992. An algorithm for the radiometric and atmospheric correction of AVHRR data in the solar reflective channels. Remote Sensing of Environment, 41, 185-195. 130 THI, A. T. P., DASILVA, J. V. & MAZLIAK, P. 1990. The role of membrane-lipids in drought resistance of plants. Bulletin De La Societe Botanique De France-Actualites Botaniques, 137, 99-114. THOMSON, B. D. & SIDDIQUE, K. H. M. 1997. Grain legume species in low rainfall Mediterranean-type environments II. Canopy development, radiation interception, and dry-matter production. Field Crops Research, 54, 189-199. THUDI, M., UPADHYAYA, H. D., RATHORE, A., GAUR, P. M., KRISHNAMURTHY, L., ROORKIWAL, M., NAYAK, S. N., CHATURVEDI, S. K., BASU, P. S., GANGARAO, N. V., FIKRE, A., KIMURTO, P., SHARMA, P. C., SHESHASHAYEE, M. S., TOBITA, S., KASHIWAGI, J., ITO, O., KILLIAN, A. & VARSHNEY, R. K. 2014. Genetic dissection of drought and heat tolerance in chickpea through genome-wide and candidate gene-based association mapping approaches. PLoS One, 9, e96758. TOKER, C. & CAGIRGAN, M. 1998. Assessment of response to drought stress of chickpea (Cicer arietinum L.) lines under rain field conditions. Turkish Journal Of Agriculture And Forestry, 22, 615-621. TOKER, C. & CANCI, H. Selection for drought and heat resistance in chickpea under terminal drought conditions. Proceedings of Fourth International Food Legumes Research Conference, 2005. 18-22. TOKER, C., KARHAN, M. & ULGER, S. 2004. Endogenous organic acid variations in different chickpea (Cicer arietinum L.) genotypes. Acta Agriculturae Scandinavica Section B-Soil And Plant Science, 54, 42-44. TOKER, C., LLUCH, C., TEJERA, N. A., SERRAJ, R. & SIDDIQUE, K. H. M. 2007. Abiotic stresses. In: YADAV, S. S., REDDEN, R. J., CHEN, W. & SHARMA, B. (eds.) Chickpea breeding and management. UK: CABI. TRETHOWAN, R. M. 2014. Defining a Genetic Ideotype for Crop Improvement. In: FLEURY, D. & WHITFORD, R. (eds.) Crop Breeding: Methods and Protocols. New York, NY: Springer New York. TRETHOWAN, R. M., MAHMOOD, T., ALI, Z., OLDACH, K. & GARCIA, A. G. 2012. Breeding wheat cultivars better adapted to conservation agriculture. Field Crops Research, 132, 76-83. TUBEROSA, R. 2011. Phenotyping drought-stressed crops: Key concepts, issues, and approaches. In: MONNEVEUX, P. & RIBAUT, J. (eds.) Drought phenotyping in crops: from theory to practice. Mexico: CGIAR Generation Challenge Programme. TUBEROSA, R. 2012. Phenotyping for drought tolerance of crops in the genomic era. Frontiers in Physiology, 3, 347. TUBEROSA, R. & SALVI, S. 2006. Genomics approaches to improve drought tolerance in crops. Trends in Plant Science, 11, 405-412. TUCKER, C. J. 1979. Red and photographic infrared linear combinations for monitoring vegetation. Remote Sensing of Environment, 8, 127-150. TURNER, N. C. 2003. Adaptation to drought: lessons from studies with chickpea. Indian Journal of Physiology, 11-17. TURNER, N. C., WRIGHT, G. C. & SIDDIQUE, K. H. H. 2001. Adaptation of grain legumes (pulses) to water-limited environments. SAN DIEGO: ELSEVIER ACADEMIC PRESS INC. UD-DIN, N., CARVER, B. F. & CLUTTER, A. C. 1992. Genetic analysis and selection for wheat yield in drought-stressed and irrigated environments. Euphytica, 62, 89-96. UNKOVICH, M. J., BALDOCK, J. & PEOPLES, M. B. 2010. Prospects and problems of simple linear models for estimating symbiotic N2 fixation by crop and pasture legumes. Plant and Soil, 329, 75-89. 131 UPADHYAYA, H. D., KASHIWAGI, J., VARSHNEY, R. K., GAUR, P. M., SAXENA, K. B., KRISHNAMURTHY, L., GOWDA, C. L. L., PUNDIR, R. P. S., CHATUVERDI, S. K., BASU, P. S. & SINGH, I. P. 2011. Phenotyping chickpeas and pigeonpeas for adaptation to drought. In: MONNEVEU, P. & RIBAU, J.-M. (eds.) Drought phenotyping in crops: from theory to practice. GCP. VAN DER MAESEN, L. J. G. 1972. Cicer L., a monograph of the genus, with special reference to the chickpea (Cicer arietinum L.), its ecology and cultivation. Proefschrift Wageningen, Veenman. VAN DER MAESEN, L. J. G. 1987. Origin, history and taxonomy of chickpea. In: SAXENA, M. C. & SINGH, K. B. (eds.) The Chickpea. Wallingford, Oxon, UK: CAB International. VAN DER MAESSEN L.J.G. 1972. Cicer L., a Monograph of the Genus With Special Reference to the Chickpea (Cicer arietinum L.), Its Ecology and Cultivation. In: BERRADA, A. F., SHIVAKUMAR, B. G. & YADURAJU, N. T. (eds.) Chickpea in cropping systems. UK: CAB Publisher. VARSHNEY, K. R., BUNYAMIN, T. A., TERESA, M., XUDONG, Z., LARISSA, D. R., AIKO, I., YING, W., WILLIAM, N., ANDREW, D. F., POORAN, M. G., CAROL, S., CHI, S., PENMETSA, R. V., CHUNYAN, X., ARVIND, K. B., WEIMING, H., PETER, W., SHANCEN, Z., JAMES, K. H., NOELIA, C.-G., JANET, A. C., HARI, D. U., RACHIT, K. S., MING-CHENG, L., MAHENDAR, T., GOWDA, C. L. L., NARENDRA, P. S., JUDITH, L., KRISHNA, K. G., JOSEFA, R., NADARAJAN, N., JAROSLAV, D., KAILASH, C. B., SARWAR, A., XUN, X., DAVID, E., GENGYUN, Z., GUENTER, K., JUAN, G., KARAM, B. S., SWAPAN, K. D., SCOTT, A. J., JUN, W., DOUGLAS, R. C., SHENG, Y., ANDREW, G. S., STEVEN, C., JONGMIN, B. & BENJAMIN, D. R. 2013a. Draft genome sequence of chickpea (Cicer arietinum) provides a resource for trait improvement. Nature Biotechnology, 31, 240. VARSHNEY, R., THUDI, M., NAYAK, S., GAUR, P., KASHIWAGI, J., KRISHNAMURTHY, L., JAGANATHAN, D., KOPPOLU, J., BOHRA, A., TRIPATHI, S., RATHORE, A., JUKANTI, A., JAYALAKSHMI, V., VEMULA, A., SINGH, S. J., YASIN, M., SHESHSHAYEE, M. S. & VISWANATHA, K. P. 2013b. Genetic dissection of drought tolerance in chickpea (Cicer arietinum L.). Theoretical and Applied Genetics, 1-18. VARSHNEY, R. K., TRIPATHI, S., RATHORE, A., JUKANTI, A. K., JAYALAKSHMI, V., VEMULA, A., SINGH, S. J., YASIN, M., SHESHSHAYEE, M. S., VISWANATHA, K. P., THUDI, M., NAYAK, S. N., GAUR, P. M., KASHIWAGI, J., KRISHNAMURTHY, L., JAGANATHAN, D., KOPPOLU, J. & BOHRA, A. 2014. Genetic dissection of drought tolerance in chickpea (Cicer arietinum L.). Theoretical And Applied Genetics, 127, 445-462. VERHULST, N., GOVAERTS, B., VERACHTER, E., CASTELLANOS-NEVARRTE, A., MEZZALAMA, M., WALL, P., CHOCOBAR, A., DECKERS, J. & SAYRE, K. 2010. Conservation agriculture, inproving soil quality for sustainable prodcution systems. In: LAL, R. & STEWART, B. A. (eds.) Advances in Soil Science: Food Security and Soil Quality. Boca Raton: CRC Press. WANG, H., LIU, C. & ZHANG, L. 2002. Water-saving agriculture in China: an overview. Advances in Agronomy, 75, 135-171. WANG, Y., XU, J., SHEN, J., LUO, Y., SCHEU, S. & KE, X. 2010. Tillage, residue burning and crop rotation alter soil fungal community and water-stable aggregation in arable fields. Soil & Tillage Research, 107, 71-79. 132 WELLS, K. 2013. Australian weather and seasons – a variety of climates [Online]. Available: http://www.australia.gov.au/about-australia/australian-story/austn-weatherand-the-seasons [Accessed 30/12/ 2016]. WHISH, J., BUTLER, G., CASTOR, M., CAWTHRAY, S., BROAD, I., CARBERRY, P., HAMMER, G., MCLEAN, G., ROUTLEY, R. & YEATES, S. 2005. Modelling the effects of row configuration on sorghum yield reliability in north-eastern Australia. Australian Journal of Agricultural Research, 56, 11-23. WHISH, J. P. M., CASTOR, P. & CARBERRY, P. S. 2007. Managing production constraints to the reliability of chickpea (Cicer arietinum L.) within marginal areas of the northern grains region of Australia. Australian Journal of Agricultural Research, 58, 396-405. WHITE, J. W., CORBETT, J. D. & DOBERMANN, A. 2002. Insufficient geographic characterization and analysis in the planning, execution and dissemination of agronomic research? Field Crops Research, 76, 45-54. WILDERMUTH, G. B., THOMAS, G. A., RADFORD, B. J., MCNAMARA, R. B. & KELLY, A. 1997. Crown rot and common root rot in wheat grown under different tillage and stubble treatments in southern Queensland, Australia. Soil & Tillage Research, 44, 211-224. WILLIAM, P. C. & SINGH, U. 1987. The chickpea: nutritional quality and the evaluation of quality in breeding programmes. In: C., S. M. & B., S. K. (eds.) The Chickpea. XU, Q. & HUANG, B. 2001. Morphological and Physiological Characteristics Associated with Heat Tolerance in Creeping Bentgrass. Crop Sci., 41, 127-133. YAN, W., HUNT, L. & KANG, M. 2001. Biplot analysis of multi-environment trial data, CAB International. YAN, W. & KANG, M. S. 2002. GGE biplot analysis: A graphical tool for breeders, geneticists, and agronomists, CRC press. YAN, W. & TINKER, N. A. 2006. Biplot analysis of multi-environment trial data: Principles and applications. Canadian Journal of Plant Science, 86, 623-645. YAN, W. & WU, H. X. 2008. Application of GGE Biplot Analysis to evaluate genotype (G), environment (E), and G× E interaction on Pinus radiata: A case study. New Zealand Journal of Forestry Science, 38, 132-142. YANG, G., LUO, X., NIE, Y. & ZHANG, X. 2014. Effects of plant density on yield and canopy micro-environment in hybrid cotton. Journal of Integrative Agriculture, 13, 2154-2163. YONG'AN, L., QUANWEN, D., ZHIGUO, C. & DEYONG, Z. 2010. Effect of drought on water use efficiency, agronomic traits and yield of spring wheat landraces and modern varieties in Northwest China. African Journal of Agricultural Research, 5, 1598-1608. ZAMAN-ALLAH, M., JENKINSON, D. M. & VADEZ, V. 2011. A conservative pattern of water use, rather than deep or profuse rooting, is critical for the terminal drought tolerance of chickpea. Journal of Experimental Botany, 62, 4239-4252. ZANELLA, F., WTANABE, T., DA SILVA LIMA, L. A. & SCHIAVINATO, M. A. 2004. Photosynthetic performance in jack bean [Canavalia ensiformis (L.) D.C.] under drought and after rehydration. Brazilian Journal of Plant Physiology, 16, 181-184. ZEWDU, T. 2002. Double cropping of forage oats and chickpea at Werata vertisol conditions, northwestern Ethiopia. In: BERRADA, A. F., SHIVAKUMAR, B. G. & YADURAJU, N. T. (eds.) Chickpea in cropping systems. UK: CAB International. ZHANG, H., PALA, M., OWEIS, T. & HARRIS, H. 2000a. Water use and water-use efficiency of chickpea and lentil in a Mediterranean environment. Crop and Pasture Science, 51, 295-304. 133 ZHANG, H., PALA, M., OWIES, T. & HARRIS, H. 2000b. Water-use efficiency of chickpea and lentil in a Mediterranean environment. Australian Journal of Agricultural Research, 51, 295-304. ZHANG, L., ZHANG, S., VAN DER WERF, W., BASTIAANS, L., LI, B. & SPIERTZ, J. H. J. 2008. Light interception and utilization in relay intercrops of wheat and cotton. Field Crops Research, 107, 29-42. ZOHARY, D. & HOPF, M. 2000. Domestication of plants in the old world: the origin and spread of cultivated plants in west Asia, Europe, and the Nile Valley, Oxford, Oxford University Press. 134