x... . as». . O-V?- . 4 . 10. .h . r 1h: .mu I , mi“ :2.le , d?.F::-€ {vaiuufi .“=&$ 5.... , .. :2. n.4,, gym .4 5.... , 4 , . . _ .44.“ r3: . :5.. 4 014 . K t~ ’54 .3...- »Ifl . . . 1.41.4 4.51:...“ .371 .1 “my... , a , ¢ 1» ur 3.1.... 0731 (ft... 4.: : .4 «if: . may? , .4... .R #32 . 1,».- ..34 54:2“... 4 . .. . . 13.. 4 variai AF .. 4.... : .z‘zm Ji. 4". .eru4 a 1:95.13.“ “Pa... €54.41 5 O .i :v as (In. £734.?“ '4; 1.4.5. 501.?! ...I m: ".30.“ “hum“! . .2342? :R tum". 4!.b4hunnrzxnu. 31$)? ‘ u. «fiat‘li‘ .v‘r‘. 4.. I) .A 3.... 3 it. .. 2.. 8.... 191...... u. ... 1.x. 3...}: Pi! 1.3.1.14, 3.3) fit: ? «2! Eur." 4.5? ..(. c I). : ‘ man: a :4“. $4.. séghtfia _ .445. :5 ll:.~ ..(..t 7. .. ‘4 v . . .. . . f. 4 an}... . 3.1, an , : . 4 . :54“. A». ..4.1. x 0 "’ LIBRARY Michigan State University This is to certify that the dissertation entitled Soybean Breeding Strategy for Pakistan: Genotype X Environment Interaction and Stability for Soybean Yield. presented by Muhammad Aslam Khan has been accepted towards fulfillment of the requirements for Ph.D. degree in Crop and Soil Sciences W&% Dr. Russell D. Freed Major professor Date Tom 30! 2002., MS U is an Affirmative Action/Equal Opportunity Institution 0- 12771 PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 6/01 c:/CIRCIDateDue.p65-p.15 SOYBEAN BREEDING STRATEGY FOR PAKISTAN: GENOTYPE x ENVIRONMENT INTERACTION AND STABILITY FOR SOYBEAN YIELD By MUHAMMAD ASLAM KHAN A DISSERTATION Submitted to Michigan State University In partial fulfillment of the requirements For the degree of DOCTOR OF PHILOSOPHY DEPARTMENT OF CROP AND SOIL SCIENCES 2002 ABSTRACT SOYBEAN BREEDING STRATEGY FOR PAKISTAN: GENO'IYPE x ENVIRON MENT INTERACTION AND STABILITY FOR SOYBEAN YIELD BY Muhammad Aslam Khan The experimental material comprising of nine soybean cultivars were planted during two seasons (spring and autumn) at four locations for three years 1992-94. Cultural operations were done when needed. Variances due to genotypes, environments and genotypes x environment interaction were highly significant for days to flowering, days to maturity, plant height, first mature pod height, 100 seed weight, number of pods per plant, oil content and yield. Whereas analysis of variance for number of seeds per pod showed non- Significant differences for genotypes, environment and their interaction, ‘NARC- lll’, ‘NARC-IV’ and “Swat-84’ showed stability, in general, for days to flowering and ‘NARC-IV’ and “Swat-84’ Showed stability in maturity. In plant height, ‘NARC-III’ and ‘NARC-IV’ performed better in good environmental conditions. Harper showed stability for pod height, whereas ‘NARC-V’ was found stable for , number of pods per plant while ‘Harper’, ‘Swat-84’ and ‘WIlliams-84’ showed moderate stability for number of pods per plant. All genotypes showed differential response and stability for 100 seed weight. ‘NARC-III’ and “NARC-IV” exhibited stability while ‘FS-85’ and ‘WIIIiam-82’ showed stability for 100-seed weight. ‘Harper’ and ‘NARC-lll' had the highest oil content. ‘NARC-III’, ‘NARC-IV’ and ‘NARC-V’ were found stable and high yielding in low rainfall areas like Fatehjang and NARC. During spring, ‘NARC-lll’ proved to be a better stable genotype for Gujranwala while ‘NARC-IV’ for Multan and ‘NARC-V’ for NARC are recommended due to their high yield potential. ‘NARC-V’ Showed wider adaptability and hence it was identified as a stable genotype on overall basis. A Plant breeding model has been described which will be very useful for the development of soybean cultivars as well as other commodities for different ecological zones in Pakistan. DEDICATION IN THE NAME OF ALLAH “THE BENEFICENT AND THE MERCIFUL” ACKNOWLEDGEMENTS Thanks GOD WHO blessed me by providing the ability and opportunity to complete my Ph.D. study. My sincere thanks goes to Pakistan Agriculture Research Council (PARC) organization for their cooperation and providing facilities for conducting research. I would like to express my deepest gratitude to Dr. Russell D. Freed, for his valuable guidance, encouragement and support during my Ph.D. program. The opportunities, which he provided, and his confidence in my abilities were greatly appreciated. I would like to express my thanks to all my committee members; Dr. Richard R. HanNood, Dr. L. Copeland and Dr. Mureri Suvedi for their excellent courses, support and helpful suggestions. My special gratitude is extended to Dr. Taylor Johnston and Dr. Doug Buhler for their support during a critical period of my education at Michigan State University. My special thanks are extended to Abdul Qadir for his help in the statistical design and analysis of data. I am also grateful to Jeff Smeenk, Jose Sanchez, Mark Bernards and fellow students for their help and friendship. My Thanks go to all professors for their support, help and encouragement. I owe special thanks to my (late) parents, brother and sisters for their support and encouragement. I would like to thank to my in-laws who always encouraged me for higher study. I also owe my thanks to Mr. and Mrs. Nawaz Khan who financially supported my study. My heartfelt thanks to my wife, Waheeda Aslam Khan, for her tireless support friendship, love, understanding and sacrifice. Thanks to my son, Mohsin Aslam Khan and my daughters, Adan Aslam Khan and Sara Aslam Khan for their unconditional support and love. vi Table LIST OF TABLES Description of Table Mean (M) and ranking (R) of 9 soybean cultivars days to flowering (DF 50%), days to maturity (DTM), plant height (pl htcm), Pod height (pdht cm), pods per plant (pods/pl), seeds per pod (sdslpod), 100-seed weight (100swt.g), oil percentage (oil %), Protein percentage (prt.%) and yield (yd kg/ha) during. autumn 1992, 1993 and 1994. Mean (M) and ranking (R) of 9 soybean cultivars days to flowering (DF 50%), days to maturity (DTM), plant height (pl htcm), Pod height (pdht cm), pods per plant (pods/pl), seeds per pod (sdslpod), 100-seed weight (100swt.g), oil percentage (oil %), Protein percentage (prt.%) and yield (yd kg/ha) during pring1992, 1993 and 1994. Analysis of variance of 9 soybean cultivars, 3 years and 4 locations for days to flowering (50%) of autumn, Spring and overall during 1992, 1993 and 1994. Stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars for days to flowering (50%) planted at 2 different locations during autumn 1992, 1993 and 1994. Stability parameters [mean, regression coefficient (b), deviation from regression (82d)] of 9 soybean cultivars for days to flowering (50%) at 3 different locations during spring 1992, 1993 and 1994.. Overall stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars, 4 locations, and 2 seasons for days to flowering (50%) during 1992, 1993 and 1994. Analysis of variance of 9 soybean cultivars, 3 years and 4 locations for days to maturity of autumn, spring and overall during 1992, 1993 and 1994. Stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars for days to maturity planted at 2 different locations during autumn 1992, 1993 and 1994. vii Page 51 52 55 56 56 57 61 62 10 11 12 13 14 15 16 17 18 Stability parameters [mean, regression coefficient (b, and deviation from regression (82d)] of 9 soybean cultivars for days to maturity planted at 2 different locations during sprin91992, 1993 and 1994. Overall stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars, 4 locations, and 2 seasons for days to maturity during 1992, 1993 and 1994. Analysis of variance of 9 soybean cultivars, 3 years and 4 locations for plant height (cm) of autumn, spring and overall during 1992, 1993 and 1994. Stability parameter (mean, regression coefficient (b) and deviation from regression (S d)] for plant height (cm) of 9 soybeans at 2 different locations during autumn 1992, 1993 and 1994. ----------- Stability parameters [mean, regression coefficient (b) and deviation from regression (82d)] of 9 soybean cultivars for plant height (cm) of 3 locations during Spring 1992, 1993 and 1994. ---- Overall stability parameters [mean, regression coefficient (b) and deviation from regression (82d)] of 9 soybean cultivars, 4 locations, and 2 seasons for plant height (cm) during 1992, 1993 and 1994. Analysis of variance of 9 soybean cultivars, 3 years and 4 locations for pod height (cm) of autumn, spring and overall during 1992, 1993 and 1994. Stability parameter (mean, regression coefficient (b) and deviation from regression (S d)] for pod height (cm) of 9 soybeans at 2 different locations during autumn 1992, 1993 and 1994. ----------- Stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] for pod height (cm) of 9 soybeans at 2 different locations during spring 1992, 1993 and 1994. ------- Overall stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars, 4 locations, and 2 seasons for pod height (cm) during 1992, 1993 and 1994. viii 62 63 66 67 67 68 71 72 72 73 19 20 21 22 23 24 25 26 27 28 Analysis of variance of 9 soybean cultivars, 3 years and 4 locations for number of pods per plant of autumn, Spring and overall during 1992, 1993 and 1994. Stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars for number of pods per plant planted at 2 different locations during autumn 1992, 1993 and 1994. Stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars for number of pods per plant planted at 3 different locations during Spring1992, 1993 and 1994. Overall stability parameters [mean, regression coefficient (b) and deviation from regression (Szdn of 9 soybean cultivars, 4 locations, and 2 seasons for number of pods per plant during 1992, 1993 and 1994. Analysis of variance of 9 soybean cultivars, 3 years and 4 locations for number of seeds per pod of autumn, spring and overall during 1992, 1993 and 1994. Analysis of variance of 9 soybean cultivars, 3 years and 4 locations for 100-Seed weight (g) of autumn, spring and overall during 1992, 1993 and 1994. Stability parameters [mean, regression coefficient (b) and deviation from regression (82d)] of 9 soybean cultivars for 100- seed weight (g) planted at 2 different locations during autumn 1992, 1993 and 1994. Stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars for 100- seed weight (g) planted at 3 different locations during spring 1992, 1993 and 1994. Overall stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars, 4 locations, and 2 seasons for 100-seed weight (9) during 1992, 1993 and 1994. Analysis of variance of 9 soybean cultivars, 3 years and 4 locations for oil content (%) of autumn, spring and overall during 1992, 1993 and 1994. 77 77 78 81 82 83 83 84 86 29 30 31 32 33 34 35 36 37 Stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars for oil content (%) planted at 2 different locations during autumn 1992, 1993 and 1994. Stability parameters [mean, regression coefficient (b) and deviation from regression (82d)] of 9 soybean cultivars for oil content (%) planted at 3 different locations during spring1992, 1993 and 1994. Overall stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars, 4 locations, and 2 seasons for oil content (%) during 1992, 1993 and 1994. Analysis of variance of 9 soybean cultivars, 3 years and 4 locations for protein content (%) of autumn, spring and overall during 1992, 1993 and 1994. Stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars for protein content (%) planted at 2 different locations during autumn 1992, 1993 and 1994. Stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars for protein content (%) planted at 3 different locations during spring 1992, 1993 and 1994. Overall stability parameters [mean, regression coefficient (b) and deviation from regression (82d)] of 9 soybean cultivars, 4 locations, and 2 seasons for protein content (%) during 1992, 1993 and 1994. Analysis of variance of 9 soybean cultivars, 3 years and 4 locations for yield (kg/ha) of autumn, spring and overall during 1992, 1993 and 1994. Stability parameters [mean, regression coefficient (b) and deviation from regression (82d)] of 9 soybean cultivars for yield (kg/ha) planted at 2 different locations during autumn 1992, 1993 and 1994. 87 87 88 90 91 91 92 95 96 38 39 Stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars for yield (kgl)) planted at 3 different locations during Spring 1992, 1993 and 1994. Overall stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars, 4 locations, and 2 seasons for yield (kg/ha) during 1992, 1993 and 1994. xi 96 97 LIST OF FIGURES Figure Description of figure Page 1 Outline of Traditional Plant Breeding Program in Pakistan -------------- 100 2 Outline of Participatory Plant Breeding Model for soybean breeding program in Pakistan 106 xii Table 1A 2A 3A 4A 5A 6A 7A 8A 9A 10A 11A 12A 13A LIST OF APPENDICES Description of Tables Structural changes in area of oilseed crops in Pakistan 1970-75 to 1998-99. Share of domestic production and import in the total availability of edible oil in Pakistan 1970-75 to 1998-99. Conventional and non-conventional source of edible oil in Pakistan during 1999. Fatty acid composition of different vegetable oils. Existing cropping system in different areas of Pakistan, ............. Proposed cropping system for soybean in different areas of Pakistan. Agro-meteorological data at NARC, Islamabad during, 1992, 1993 and 1994. Agro-meteorological data at Fatehjang during, 1992, 1993 and 1994. Agro-meteorological data at Gujranwala during, 1992, 1993 and 1994. Agro-meteorological data at Multan during, 1992, 1993 and 1994.—- Estimated Potential Areas for the cultivation of Oilseed Crops in Pakistan. Map of Pakistan. Area, production and yield (kg/ha) of soybean in Pakistan, 1980- 99 xiii Page 120 121 122 122 123 123 124 125 126 127 128 129 130 Abbreviation AARI ARI b C CM CV DF et al. F ath F 8-85 Guj kg/ha Max Min Mul NARC NS ORI % List of Abbreviations Word of sentence Ayub Agricultural Research Institute Agricultural Research Institute Regression coefficient Centighrade Centimeter Coefficient of variation Degree of freedom and others F atehjang Fasialabad-85 Gujranwala Kilogram per hectare Maximum Minimum Millimeter Multan National Agricultural Research Centre Non-significant Oilseed Research Institute Percent Coefficient of determination xiv Sd Std.error Deviation from regression Standard error XV LIST OF CONTENTS DEDICATION ACKNOWLEDGEMENT LIST OF TABLES LIST OF FIGURES LIST OF APPENDICES LIST OF ABREVIATIONS. LIST OF CONTENTS I. INTRODUCTION. II REVIEW OF LITERATURE. -Genotype x Environment interaction and stability. —Participatory plant breeding. III MATERIALS AND METHODS -Experimental materials and methods. -Data recording procedure. -Statistical procedure -Genotype x Environment interaction model -Stability model - Mean -Regression coefficient -Deviation from regression -Genotypic Coefficient of Variation -Participatory Plant breeding model -Priority setting -Sustainable funding -Participatory planning and implementation IV RESULTS AND DISCUSSION -Genotype x Environment interaction and stability -Days to flowering -Days to maturity -Plant height (cm) -Pod height (cm) -Number of pods per plant -Number of seeds per pod - l OO—seed weight xvi Page iv vii xii xiii xiv xvi 38 39 41 43 43 43 43 44 44 44 45 45 45 45 47 48 50 58 64 69 74 79 79 -Oil content percentage -Protein content percentage -Yield (kg/ha) Social challenges in plant breeding — Traditional plant breeding model - Participatory plant breeding model V SUMMARY LITERATURE CITED APPENDIX xvii 85 89 93 98 99 10] 107 111 119 INTRODUCTION CHAPTER—I INTRODUCTION Agriculture is the largest sector of the Pakistan economy and contributes more than 24.5 percent to the Gross Domestic Production (Pakistan Economic Survey, 1998-99). Most of the population directly or indirectly depends on this sector. Agriculture employs about half of the total employed labor force and is the main source of foreign exchange earnings. The performance of agriculture depends largely on the vagaries of nature. Despite the best efforts of the farmers, climatic conditions adversely affect the output of crops. In fact, the development efforts in the agriculture sector in the past had primarily focused on production and development of cereal crops like wheat and rice and cash crops like cotton. These efforts have provided rich dividends in the form of self-sufficiency and production of exportable surplus. As a result, the land area under cereal and cash crops increased while it decreased under oilseed crops (Table 1A). The worsening situation of a decrease in production of edible oil now demands a major shift with endeavors to develop varieties of different oilseed crops with high yield potential under different environmental conditions. Pakistan is chronically deficient in the production of edible oil; so much so that 72 percent of the country’s requirements are met through imports costing huge amounts in foreign exchange (Table 2A). One disturbing aspect of this crisis is an annual increase of approximately 50,000 tons in the gap between consumption and domestic production (FAO 1999). It is a matter of great concern that efforts made so far to enhance domestic production of edible oil have not had any significant impact. Edible oil is second only to petroleum among commodities being imported into Pakistan and requires a large amount of 2 foreign exchange. During 1998-1999, the total domestic production of edible oil was 532 thousand tons, of which about 72 percent was obtained from cotton, while Brassica sp. contributed 14 percent and the remaining production was from other crops like sesame, sunflower, soybean, safflower and maize (Table 3A). The factors contributing to the increase in consumption of about 50,000 tons per year are as follows: 0 Population increases 0 Per capita consumption increase 0 Increasing urbanization 0 Higher family incomes leading to a better stande of living 0 Decreased per capita availability of animal fat and its higher cost Soybean is an excellent crop for edible oil and its seeds contain 18-20% oil, 40-42% protein, 19-25% carbohydrate, 4-5% fiber, 5% ash, and 80-85% unsaturated fatty acid plus vitamin B and no cholesterol. Thus, soybean would provide not only good quality oils but also much needed protein for Pakistan. Animal meat is very expensive and is not available in good quality. Therefore, soybean will not only be a source of edible oil but also provide a source of quality protein, which an ordinary farmer can afford. The total area under soybean production in the world is about 69.4 million ha yielding 152.6 million tons (Soy and oilseed Blue book, 1999), which signifies its importance and popularity as food and a versatile crop. Of this total, the United States contributes 47%, Brazil 20%, Argentina 11%, China 10%, India 3%, Canada 2%, and Paraguay 2%, The European Union contributes 1% and other countries, 4%. Soybean has a lot of potential and can be successfully grown in Pakistan. The existing cropping system (Table 5A) in the cotton area is cotton-wheat-cotton (Akhtar et a1. 1986) while in the rice area, it is rice-wheat-rice-wheat (Byerlee et al., 1984) and in the rain- fed area wheat- sorghum - fallow or wheat - fallow - wheat (Byerlee et al., 1984; Hobbs et al., 1983; Hobbs et al., 1984; Riaz et al., 1983;). The cotton crop yield has been affected by the existing cropping system due to continuous planting of the same crops without following any rotation. In the cotton area, the last picking is done during mid January, and it will not be economical to plant wheat at that time. But most farmers’ complete cotton picking occurs during the 1St week of December, with wheat planting done during the 2nd week of December. This is not as profitable as if wheat planting is done between October 15 to November 15, which is the optimum time to get the maximum yield of wheat. But the farmers follow the existing system because there is no other choice except wheat. Similarly, in rice area, the land is wet and not ready for wheat planting at the proper time so the planting of wheat is delayed, which affects its production. Results of the experiments conducted under different ecological conditions indicate that soybean can be productive and can fit into a cotton, rice and rain-fed cropping system (Table 6A). In rainfed areas, there is ample of rain (Table 7A — 10A), which can be utilized by planting soybean in the fallow area. Soybeans can fit into new cropping systems in the cotton, rice and in the rainfed area because 30%, 30% and 25% remains fallow in these areas respectively due to one reason or the other (Table 11A). As a leguminous crop, soybean will improve the soil fertility (Porter et al. 1997). By planting soybeans in rotation, soil erosion, diseases and insect damage will be minimized. Soybeans will also provide additional income to the farmers. The commercial cultivation of soybean in Pakistan started in the early 1970. Prior to that time, soybeans were practically unknown in the plains of Punjab and Sindh (Table 12A). However, some local black and chocolate colored varieties of pulses now identified as soybean have been grown from time immemorial in the hills of northern parts of Pakistan i.e., Hazara, Azad Kashmir, Swat, Dir and Kurram Agency. To improve the prevailing varieties and to introduce new ones, some research was started in 1965 in North Western Frontier Province (NWFP). Since the establishment of Oilseeds Coordinated Program at Pakistan Agricultural Research Council (PARC), Islamabad in 1977), extensive research work for the development of early maturing and high yielding varieties is in progress at the National Agricultural Research Center (NARC), Islamabad, The Ayub Agricultural Research Institute (AARI), Faisalabad and the Agricultural Research Institute (ARI) in Tandojam, Tamab (Peshawar), and Sariab (Quetta). During the past years, efforts have been made by agricultural scientists to introduce soybean commercially but without much success. Expansion of the area planted in conventional oilseed crops is difficult. Still, new varieties of soybean can be developed to fit in the cropping system in the cotton, rice and rain fed areas. During 1970-71, soybean was planted on an area of 2460 hectares, which declined to 1660 hectares in 1976-77. From 1977-78 onward area increased slowly to 5980 hectares till 1986-87, then it fluctuated again in 1998-1999. The production from 1980-1999 is given in Table 13A where fluctuation can be seen for each year. Recent research revealed that due to lack of adapted high yielding and early maturing soybean varieties, the actual national average is 1/2 of the potential yield. Therefore, to increase the production per unit area, new high yielding and early maturing 5 varieties need to be developed which are suitable for different ecological regions. The national average production (Table 13A) is very low at this stage. The study of genotype x environment (G x E) interaction has assumed great importance in variety testing programs because the yield performance of a genotype is the result of interaction with the genotype and environment. Environmental factors such as rainfall, temperature and soil fertility play an important role in varietal performance and yield. Therefore, the release of a genotype with consistent performance over a wide range of environments should lead to stability in production. A measure of the relative yield stability of the soybean varieties under a wide range of environmental conditions is essential for determining efficiency in a variety evaluation program. Genotype x environment interaction studies on soybean has not been done in Pakistan. This initial study will help the development of suitable cultivars for different ecological conditions without disturbing existing cropping systems. lnforrnation from the study will determine potential parents that will be used in the genetic improvement program. As a result, scientists should be able to develop suitable cultivars for different environments. The crop area under soybean will be increased which will narrow the production and consumption gap. Substantial work on this aspect has been done in other countries. In most of these studies (Baihaki et al., (soybean) 1976; Baker, (wheat) 1969; Borojevic and William, (wheat) 1982; Brennan and Byth, (wheat) 1979) data from several locations have been used for determining the significance of G x E interaction. However, in the studies of wheat and cluster beans (Luthra et al., 1974; Saini et al., (1977) involving single location experiments, the different environments were created by using different dates of planting, spacing, fertilizer rates, irrigation levels, etc. Participatory plant breeding (PPB) is defined as a breeding approach that involves a close relationship between farmers and researchers to bring improvement within a plant species. In this dissertation, PPB will refer to the full scope of activities associated with plant genetic improvement which include identification of breeding objectives, generating genetic variability, selecting within variable populations to develop experimental lines, evaluating experimental lines termed “participator variety selection” (PVS), variety release, popularization and seed multiplication activities. Farmers’ participatory plant breeding programs have several important goals, the most important of which is to increase crop production in the farmers’ fields. Increasing the income of the farmer through development and adoption of high yielding improved soybean cultivars is important for Pakistan. This study will develop a new model for plant breeders to set their research agenda. It will address the important issues of today: 0 The research-extension linkage o Participatory Planning and implementation 0 Sustainable financing 0 Priority setting This model will outline how priorities can be set for the breeding program. This model will be applicable to all crops in both the developed and developing world. REVIEW OF LITERATURE CHAPTER-II REVIEW OF LITERATURE GENOTYPE-ENVIRONMENT (G x E) INTERACTION AND STABILITY: Genotype-environment (G x E) interactions have concerned plant breeders for many years. Different procedures have been used to characterize individual varieties for expression in varying environmental conditions. Performance tests are conducted at a number of locations and repeated over several years to find how cultivar response varies to different locations and different climatic conditions. The experiments conducted at a single location and in a single year have a limited utility. Therefore, performance tests over several years and locations are important for the development of cultivars with wide adaptability. Yates and Cochran (1938) proposed a regression model to evaluate the adaptability of a variety over diverse environments. A variety or genotype is considered to be more stable if it has high yielding ability when growing in diverse environments. They used the mean performance of all genotypes grown in an environment as a suitable index of its productivity. The performance of each genotype was plotted against this index for each environment and a simple linear regression was fitted by least squares to evaluate the genotype’s response, the mean of the regression slope being one. Jensen and Neal (1952) suggested that genotype x environment interactions would be reduced by using multiline varieties of oats rather than pure line varieties because such multiline varieties would possess greater stability of production, broader adaptation to the environment and greater disease resistance. Although it is possible through population genetics and biometry to develop breeding programs for crop improvement, the plant breeder must select on the basis of an 9 ever-changing environment. Grafius (1956) studied the components of yield in oats. He stated that yield is not only a ratio but it is also a product. He also concluded that recurrent selection for a ratio, without strong positive correlation between the heritable numbers of the ratio, was likely to be firtile. Plaisted and Peterson (1959) computed an analysis of variance for every pair of genotypes to estimate the interaction variance for every combination of two genotypes. The mean of the interaction variance obtained for each genotype was used as an indicator of the contribution of that genotype to the total genotype x environment interaction. Plant breeders and agronornists are interested in the potential performance of crop varieties over a wide range of ecological conditions. In a study by Finlay and Wilkinson (1963), varieties from particular geographic regions of the world showed a similar type of adaptation, which provides a useful basis for plant introduction. For each variety, a linear regression of yield on the mean yield of all varieties for each site and season was computed to measure variety adaptation. The mean yield of all varieties for each site and season provided a quantitative grading of the environment and from the analysis, Wilkinson concluded that varieties adapted to good or poor seasons and those showing general adaptability can be identified. It is well known that phenotype reflects non-genetic as well as genetic influences on development and that the effects of genotype and environment are not independent. The phenotypic response to a change in environments is not the same for all genotypes. The main objective of most breeding programs is to select genotypes that will perform well in a broad spectrum of environments or varieties highly adapted for a specific environment. A variety that is superior only in a specific environment has less value than widely adapted 10 cultivars. Comstock and Moll (1963) described the variances which are pertinent to plant breeding problems are those associated with variety x year, variety x location and variety x year x location interactions. Estimates of the magnitudes of these variances relative to each other and relative to the error or plot to plot environmental variance are necessary for the application of existing quantitative genetic theory to plant breeding. They have shown statistically the effect of large G x E interaction in reducing progress from the selection. Allard and Bradshaw (1964) classified the variation of the genotype x environment in two types: predictable and unpredictable. Significant variety x location or variety x year interaction suggested that the appropriate breeding program should allow for the development of a number of varieties, each particularly adapted to one of the special environments. Variety x year interaction is very different from variety x location interaction because year-to-year fluctuation cannot be predicted in advance. They concluded that genetic diversity either as heterozygous or as a mixture of different genotypes often leads to stability under varying environmental conditions. The occurrence of G x E interaction provided a major challenge in obtaining a fuller understanding of the genetic control of variability. Phenotype is the product of genotype and its environment. Finlay and Wilkinson (1963) developed a statistical technique to compare the yield performance of a set of cereal varieties grown at several locations for several seasons. This involves computing for each variety the regression of individual yield on the mean yield of all varieties for each site and season. For the varieties and sites tested the regressions had a high degree of linearity and were used as measures of the adaptability of the varieties. Similar techniques yielding similar results were reported by Yates and Cochran (1938). Breese (1969) conducted research in grasses to measure the significance of 11 G x E interactions. His study involved computing for each variety the regression of each individual on the mean yield of all varieties for each site and season. The wide range of genetic material showed marked interaction with contrasting climatic, edaphic and management conditions. The major part of this interaction could be explained by differences between responses as estimated by regression. The results have been used to demonstrate that this method can be a powerful means of predicting relative performance of populations and their hybrids over seasons, years and locations. Green et al., (1972) emphasized the need to develop high yielding cultivars adapted to a specific production environment. The existence of significant interactions between genotypes and the environment supports the need of tailoring genotypes to fit specific environments. Selecting the genotypes possessing the phenotype that permits the best exploitation of the yield potential of a specific production environment is a major objective of plant-breeding programs. To accomplish this objective, the breeder must determine empirically or mechanistically the characteristics and nature of the ideotype so that appropriate germplasm manipulation could be made. Baihaki et al., (1976) conducted a study to determine the relationship of G x E interaction to yield level (high, medium, low) in preliminary yield test material. The regression of stability parameter estimate (b value) on mean yield of the lines was positive and highly significant. In general, the medium yield group of lines was most stable and the low yielding group was the least stable. They concluded that a single environment preliminary yield test could be used without risk of discarding outstanding - lines. Specht and Williams (1978) observed in a temperate, rain-fed climate that soybean lines used in diverse regions were not quite adapted and were not very yield stable in terms 12 of high temperature response. Since temperature in July is very high, breeders must keep in mind the phenotype most likely to elicit the best yield response in such an environment. Critical consideration of these interactions between genotype and the environment permits the breeder to make some mechanistic determinations about the problem comprising the ideotype and supplement the subsequent empirical determination necessary to confirm or deny tentative judgements concerning the nature of the ideotype. The joint expression of non-genetic and genetic factors is reflected by phenotypic performance. The genotype may respond differently to various environmental factors or groups of environmental factors. Baker (1969, for wheat;) and Abou-El-Fittouh et al. (1969, for cotton) found in their studies that the magnitude of the G x E interaction variance has been found to be larger than that for genotypes, making it difficult to recognize small actual differences in yield. Brennan and Byth (1979) studied G x E interaction of widely adapted wheat genotypes. They concluded that greater selection differentials were found in environments when selection was practiced for high mean yield across all environments when the yield of each cultivar in each environment was expressed as a percentage of the environment mean yield. Since the expression of a character is dependent on the interaction of the genotypes and the environment, the stability of performance in the context of different environments is an important criterion in discriminating varieties. Sharma et al., (1980) conducted research on 28 genotypes of soybeans in kharif (autumn) and spring seasons for 2 years to evaluate level and stability of performance for days to first flowering, days to maturity, and seed yield per plant. They found that the mean differences among the genotypes were highly significant for all the characters. Highly significant genotype x environment interactions were observed for all 3 characters suggesting that the characters were significantly 13 influenced by changes in the environment. The variances due to environment (linear) were significantly different for all the 3 characters, indicating that the response to environments was genetically controlled. The genotype x enviromnent interaction (linear) component of variation for stability was also highly significant for days to first flowering and non- significant for days to maturity and seed yield, which indicated the differential response of the genotypes to different environments. Genotype x environment interaction is important to plant breeders because of the confounding effects it introduces in comparisons among genotypes tested in different environments. Saeed and Francis (1984) conducted research in grain sorghum to determine the relative contribution of several weather variables during various plant growth stages to variation in environment and genotype x environment interaction of grain sorghum genotypes in different maturity groups. They found that variation in weather factors contributed more to G x E interaction for yield of the late maturing genotypes than for the early and medium maturing genotypes. Evaluation of genotypes for consistency of performance in different environments is important in plant breeding programs. The relative performance of genotypes often changes from one environment to another. The occurrence of a large genotype x environment interaction poses a major problem of relating phenotypic performance to genetic constitution and makes it difficult to decide which genotypes should be selected. It is important to understand the nature of G x E interaction to make testing and ultimately selection of genotypes, more efficient. Saeed et a1. (1984) studied the effects of genotype maturity on G x E interaction of sorghum for grain yield and yield components. The genotypes were of three maturity groups; early, medium and late and evaluated in 48 environments in 2 years. They suggested that a reliable evaluation of relative genotype performance could be made if 14 the genotypes under test do not include a wide range of maturities. Testing at more locations should be done rather than testing in more years. For a desired level of precision, the amount of testing required for a set of genotypes with a wide range of maturity and long- growth duration is greater than that required for genotypes with a narrow range of maturity and short-growth duration. Grain yield is complex and is determined by various qualitative and quantitative characters, which are greatly affected by the enviromnent. Thus, it becomes very important to ascertain the extent of influence of changing environments on grain yield and yield components before recommending a variety for commercial cultivation. Malik and Rajput (1984) conducted research to study genotype x environment interaction on wheat. They concluded that prior to variety release, yield potential and adaptability should be ascertained under different environmental conditions. Pathak and Nema (1984) observed that shorter growth periods between flowering and maturity as well as between pod set and maturity were associated with high yield in an early sown crop while longer growth periods between the beginning and end of flowering and between flowering and maturity were associated with increased yield in a late sown crOp. Genotype-environment interaction is important in evaluating variety adaptation, selecting parents for breeding programs according to the environments, and improving genotypes with improved adaptability. Mariani and Manmana (1986) concluded from the combination of results from different varieties, years and locations that all the regression models applied showed similar efficiency values and analogous stability parameters for each variety. In addition, they also verified by multi-phase regression models whether a specific response to poor, average or rich environments arose for a variety. 15 The presence of G x E interaction creates difficulties in assessing the performance of different cultivars. This is especially true in food legumes, which are often cultivated under adverse agro-climatic and management conditions. Therefore, it is essential to identify the cultivars which have high yield potential and have stability over a wide range of environments before they are released. Singh and Chaudhry (1985) computed stability analysis for protein and oil contents in different soybean lines in artificial environments. Significant environment and genotype interaction indicated variable response of varieties to a wide range of environments. Similarly, significance of the linear component of environment and G x E interaction indicated the predictability of regression coefficient pertaining to various genotypes on environmental means. They concluded that genotypic differences and G x E interaction were significant for oil content only. The role of genotype-environment interaction is well known in crop plants. A genotype grown in different environments may have different responses. Similarly, when hybrid populations are grown in different environments, inconsistent phenotypes may appear due to the interaction of genotype and environment. A hybrid population of a self- pollinated crop advanced for several generations under different environments would be very usefirl in the study of evolution, adaptation, breeding parameters, and genetic variation of the crop. Selection can be carried out in early generation and desired characters can be selected to identify that plants express their potential characteristics when hybrid I populations are advanced under favorable conditions [(Iman and Allard, (1965), Goth (1955) and Adair and Jones (1946)]. Chan et al. (1986) studied the frequency distribution of agronomic characters in the F2 generation of soybean. They concluded that the standard deviation of plant height, nodes on main stem, days to flowering and days to maturity tended 16 to increase from south towards north and suggested that that some differentiation in the F2 populations under various locations was involved. Genotype x environment interaction plays an important role on the stability performance of genotypes. If this interaction is high, genotypes become highly adapted to the specific environments. However, if the interaction is low, the genotype performs unifome over a wide range of environments. Thus, selecting genotypes for performance stability under varying environmental conditions has become an essential part of any breeding program. In order to know the effect of interaction, a number of statistical and biometrical-genetical approaches have been developed by different authors (Finlay and Wilkinson, 1963; Ebberhart and Russell, 1966; Perkins and Jinks, 1968). Konwar and Talukdar (1986) studied soybeans genotypes in different environments to evaluate the stability of performance of genotypes and to identify the stable genotypes with high yielding ability. Sufficient G x E interaction was exhibited by the genotypes for all the characters. However, the characters differed in regards to the contribution of linear and non-linear components of G x E interaction variance. The genotype Bragg exhibited average stability for seed yield per plant followed, by DS73-16 and Kalitur; JS72-375 for number of pods per plant and number of clusters per plant; and PK327 for 100-seed weight exhibited above average stability which can be utilized in a breeding program. Genotype-environment interaction is important in selecting parents for genetic improvement programs in different environments. Mariani and Manmana (1986) studied the combination of results from several data sets to assess the adaptability of Italian wheat varieties and to verify if stability parameters of varieties planted in some of the environments had the same reliability. Their results revealed that multi-phase regression l7 models applied to the three sets were not significantly different from linear regression; this was used to measure the average sensitivity over all environments for each variety. They also verified by multi-phase regression models whether a specific response to poor, average or rich environments arose for a variety. Stability of performance across environments is considered essential for the release of soybean genotypes with improved oil quality. Recent advances in the development of soybean genotypes with higher oleic acid and lower linolenic acid percentage may be accompanied by changes in genotype x environmental interaction of the selected material. Carver et al., (1986) conducted research on soybean to evaluate stability of unsaturated fatty acid composition in lines derived from the original population and from the third and sixth cycles of recurrent selection for high oleic acid. The regression analysis showed that genotype x environment interaction for each unsaturated fatty acid could be attributed to differences among genotypes in their linear responses to changes in the environmental mean. The stability analysis revealed a higher proportion of selected lines sensitive to environmental variation in oleic and linoleic acid percentages. Sen and Mukherjee (1986) suggested autumn and late autumn to be relatively more favorable than the rabi and late rabi. They also observed no association between mean yield and stability performance. Molari et al., (1987) reported difficulties in identifying plants in the progenies of a soybean cross and considered it to be due to genotype x environment interaction. They suggested that selection for yield should be based on characters of simple genetic control such as pods per plant. Selection response was estimated for each generation from a comparison of the selected population to the whole population. Early maturity showed a highly significant selection response for seed yield. There was a significant positive response only in the F 5 and this is attributed to a year effect. 18 Screening of advanced strains for environmental adaptability is very inrportant in India where barley is cultivated over varying soil moisture conditions. Verma et al., (1987) conducted research to evaluate elite lines of barley for yield stability under different environmental conditions. According to pooled analysis there were significant differences among the strains. The significant genotype by environment component indicated real differences in varietal performance over environments. Variance due to pooled deviations was highly significant, indicating that differences in stability were due to the deviation from linear regression and not to the inferences in the slope line. Odendaal and Deventer (1987) studied genotype x environment interaction that plays an important role for grain yield and protein percentage on soybean cultivars planted at 9 different sites for 3 years. The genetic variance components for yield and protein percentage amounted to 10% and 24% of the total phenotypic variability respectively. The variance components for the cultivators x location and cultivar x year interactions were small for both trials. The second—order interactions were substantial, amounting to 33% and 12% for yield and protein percentage respectively. The error component accounted for 48% and 55% of the respective total phenotypic variances. Hildebrand et al., (1988) reported that large differences in oil stability analysis were seen in 43 soybean lines. Differences in seed filling and maturing may have affected oil stability. Development of cultivars specifically adapted to later planting dates commonly associated with double-crop production has been suggested as a means to expand double- crop in the area. Raymer and Bernard (1988) conducted research on soybean cultivars to determine if currently used soybean cultivars differ in adaptation to late planting and if any specific traits are related to improved performance under late-planted conditions. Their l9 research was conducted on 16 cultivars for 3 years and 2 dates of planting. Cultivars by planting date interactions were found to be significant for days to maturity, height at maturity, seed quality and seed mottling, but not for yield, days to flowering, height at flowering, lodging and weight per 100 seeds. All cultivars suffered substantial and similar yield reductions when planted late. Phenotypic correlation coefficients of cultivar performance between the two planting dates were positive and highly significant for all plant traits. The relationship of yield with various plant traits varied greatly from year to year and no differences in these relationships were observed between the two planting dates. They concluded that the lack of a strong cultivar by plant date interaction for yield and the lack of any strong associations of specific plant characteristics with yield in late planted environments imply that testing in a conventional early-planted environment will be effective in identifying lines that perform well in either full-season or double-crop environments. To make a rational decision on whether a special selection program is necessary for improving soybean grain yield in an intercrop, information is needed on the effect of intercropping on the expression of genetic variability and genetic parameters, nature and magnitude of association of different traits, and direct and indirect effects of various traits on seed yield. Sharma and Mehta (1988) studied the effect of cropping systems on genetic variability and component analysis in soybean. According to their results, the correlation coefficients between traits were found to differ, both in nature and magnitude, between monoculture and inter-cropping. In monoculture, seed size, harvest index and oil percentage was positively related with seed yield. By contrast, plant height, branches per plant, pods per plant and pod clusters per plant, besides lOO-seed weight and harvest index, were correlated with seed yield under inter-cropping. Seed size followed by 20 branches per plant appeared to be the important characters for higher seed yield selection under sole cropping. Yield improvement in inter-cropping was associated with increased harvest index. Soybean is basically a rainy season crop in India but it can be successfirlly grown during rabi and summer seasons. The area of soybean in India is increasing very rapidly to bridge the gap between demand and supply of edible oil and protein. Therefore, a study was conducted to evaluate the adaptation of promising soybean cultivars to different seasons. Patil et al., (1989) found significant differences between environments and nonsignificant differences between 25 genotypes of soybean for seed yield. Mean square (MS) for both environment and genotype were observed for days to flowering, days to maturity, plant height and pods per plant when tested against pooled error. Highly significant MS for environment (linear) showed wide differences between environments and had considerable influence on all the characters. Pooled deviations were significant for all the characters indicating significant differences with respect to stability between them. Five genotypes of soybean were found to be comparatively stable for days to flowering whereas for days to maturity. None of the genotypes were significant for date of maturing. Five soybean genotypes had above average stability for plant height. Based on regression coefficient and deviation from regression values, six genotypes showed stability for number of pods per plant. As in India, soybean is also a rainy season crop in Pakistan but it can also be grown in irrigated areas in Pakistan due to availability of land and irrigation facilities without disturbing the existing cropping system. However, the area under soybean has not been increased in Pakistan due to non-availability of high yielding cultivars suitable in different 21 ecological zones. Patil and Narkhede (1989) conducted research on phenotypic stability of yield, seeds/pod and 100 seed weight in black gram to collect information for launching a dynamic and efficient breeding program. Pooled analysis of variance revealed the presence of genetic variability as well as variable environments for the characters under study. The G x E interaction including environmental linear effects was found to be significant. Highly significant differences were found due to environment for all traits. The linear component of G x E interaction was highly significant for seed yield, seeds/pod and lOO-seed weight. They concluded that the prediction for most of the genotypes appeared to be feasible for most characters. Significant pooled deviation for all three characters suggested that the genotypes differed considerably with respect to their stability for these characters. Yadar and Kumar (1983) also reported the linear and non-linear components of G x E interaction for seed boldness in black gram. Clark and Snyder (1989) reported a Significant interaction between year and cultivar for seed oil content in different soybean cultivars. Cultivar differences vary by year, and year differences vary by cultivar. The growing conditions were different for each year, but the cultivars did not respond the same to changes in growing conditions. This indicates that cultivars may have individual responses to environmental conditions that influence oil content. The existence of significant G x E interaction creates difficulty in genetic analysis in several ways, such as by confounding estimates of genetic parameters and statistics, and by complicating selection and testing strategies. Such interactions reflect differences in adaptation, which may be exploited by selection and testing strategies. There is conflict between breeding for broad adaptation (minimizing interaction) and for specific adaptation (emphasizing favorable interactions). However, any objective decision requires a full 22 understanding of the nature of G x E interaction. Kroonenberg and Basford (1989) investigated multiple attributes (seed yield, plant height, and seed protein percentage and oil percentage) in different soybean genotypes at different locations. Standard analysis of variance indicated that many significant differences and interactions exist, but did not give specific information about the response patterns. However, when results of seed quality were compared with emergence in field planting, sometimes inconsistent results have been found. Some inconsistencies among investigations may be due to use of different procedures; however, much of the variation has been attributed to variations in the planting environment. This suggestion implies recognition that different seed lot characteristics may be better related to performance in different field situations, and it has been proposed that seed quality tests should simulate the stresses imposed by different environments (Schoorel, 1957; Woodstock, 1973). Ferris and Baker (1990) reported G x E interaction stability emergence results for five non-flooded soil environments that indicated significant linear and quadratic interaction between seed and environment. From their results, they concluded that the seed quality test, which is best for predicting soybean emergence, could vary with seedbed conditions. The interaction is usually present whether the varieties are pure lines, single cross or any other material with which the breeder may be working. Qari et al., (1990) conducted a study on ten barley varieties to evaluate for stability over nine locations for two years. Combined analysis of variance revealed significant differences among varieties for yield performance. The G x E interactions were, however, of the non—linear type because variety x environment (linear) was non-significant against pooled deviation. The non-significance of variety x environment (linear) reveals a lack of genetic difference among varieties for their 23 response to varying environments. The variance due to pooled deviation was highly significant, indicating that an important component for difference in stability was due to deviation from linear regression only. Plant breeders continuously try to find ways to increase the efficiency of selection for seed yield. An important factor in the development of cultivars is yield testing at different locations. Rosielle and Hamblen (1981) suggested that the most desirable approach for plant breeders would be to choose testing sites that are representative of the production areas. Whitehead and Allen (1990) conducted research on soybean to test the genotypes that have superior yield potential in both low and high stress edaphic conditions. Their results indicated that low-stress environments commonly used in soybean breeding programs should provide high probabilities for selecting genotypes that have superior yield potential in both low-and high stress edaphic conditions. Regression coefficients for line mean yields (regressed on site mean were generally highest for the superior lines (b> 1.0) but the deviation from regression was similar to the non-superiors lines. The ranges for maturity and plant height were similar for both groups indicating that selection for superior lines would not have led to extremes for either trait. Harer and Shah (1991) studied stability for seed yield in some soybean varieties during kharif and rabi 1986, 1987 and summer season of 1988. They found that the mean differences among genotypes were highly significant which reveals the presence of genetic variability among the genotypes. Highly significant mean squares due to environments and genotype x environments interaction revealed that the expression of genotypes varied in different environments. The partitioning of G x E interaction into linear and non-linear portions indicated that the mean squares due to linear components played an important role 24 in total G x E interaction. The excess of the linear component revealed that the prediction of performance in different environments would be possible. The stability parameters revealed that all the varieties showed significant linear and non-significant non-linear components. Mayers et al. (1991) conducted research on adaptation of soybean in three tropical dry season environments to examine genotype and environmental effects on growth and seed yield per plant. They concluded that dry matter (DM) at maturity increased exponentially with crop duration and was greater with later maturing genotypes. Across environments, thermal time provided better description of DM accumulation than did crop duration, indicating direct effects of temperature on growth rates. Among genotypes, the relationship between seed yield and DM production was strongly linear, implying that under the wide spacing of the study, DM production was the main basis of genotype differences in seed yield. Among environmental means, the relationship was weaker and curvilinear; suggesting those environmental effects of vegetative growth were not necessarily reflected in seed yield. They also found that where phototherrnal regimes delayed flowering and maturity, vegetative growth was excessive, and the harvest index (HI) smaller. HI was also smaller where flowering coincided with cool night temperatures (.om> O NNNF O F.OF. O OON N NF N N O NN N F. O N O OFF O NO ONO; l. NOOF N NOF. F. NON N OF F N N ON N F. F. NO F NNF O NO F.N-FO;O F OOON F OOF. N O.FN N OF N N F ON N F. F NO N NNF F. NO >-om_-Om H > o .26. 552% FO.O FO FOOOFFEOO .. §c®.N OF I I I I 0 X m X > cum.m—. w I I I I O X m cam. rm N I I I o m X > ..N.ONF.O F - - - - FOO 88% :2. NO :ON NN :3. OF > x ._ x o :3 N ..F.. OF :ON O o x 4 OS OF ..N.N OF :NOF OF 0 x > :NONN N ..F.NO N :NNNF O FOO 55:30 ON O ..O.NOF F. OR N ._ x > ..F.OFF N :NOOOF N 1.8 F :V OOORFS :OON N ..N...OF N :N. FN N 8 am; 2...; n. “5 2...; u. “a 2...; O “a OOFBON ._._o oz_EN Est? .39 new 89 .33 acts—O =Em>o new mctam 62.2%.. .6 3°08 accuse: 9 9:6 .2 25:80. F. can NEON m .2923 OmegoO a No 62.6th, .6 £9.92 .m min... Table 4. Stability parameters [mean, regression coefficient (b) and deviation from regression 82d)] of 9 soybean cultivars for days to flowering (50%) planted at 2 different locations during autumn 1992, 1993 and 1994. Mean 33 Std. error 0.23 CV 4.56 Table 5. Stability parameters [mean, regression coefficient (b), deviation from regression (Szd)] of 9 soybean cultivars for days to flowering (50%) at 3 different locations during spring 1992, 1993 and 1994. 1:29 .27 Mean 61 Std.error 0.24 CV 2.25 56 Table 6. Overall stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars, 4 locations, and 2 seasons for days to flowering (50%) during 1992, 1993 and 1994. -84 47 . 48 0.53 51 0.65 l 50 -0.51 NARCI-V 53 -1.15 52 1.35 A -84 51 -1.32 WEBE 44 -0.11 51 0.98 Mean 50 Std.error 0.61 CV 2.98 57 Days to maturity: Days to maturity is a very important variable for selecting genotypes for the different agroecological zones and different cropping systems in Pakistan. Several maturity types are needed to fit the different cropping systems. However, short season cultivars fit better into the different cropping systems in Pakistan. The pooled analysis of variance for days to maturity at 4 locations in 3 years (1992-94) during spring and autumn and overall is presented in Table 7. Highly significant differences among cultivars, location, years and their interactions were found indicating the presence of variability and sensitivity of cultivars to different environments. Similar results were observed by Cianzio et al., (1991). Estimated variance due to environment (linear) was much greater than the estimate of genotype effect, hence made a greater contribution to the total estimated variance (Naidu et al., 1991). Days to maturity is highly affected by the season: during the autumn, the cr0p takes less time to mature. The cultivar x location, cultivar x year, and cultivar x year x location interactions were highly significant; they were also significant when tested against pooled deviation, which revealed that there are real differences among varieties for their regression on the environmental index. However, the present investigation clearly indicates that days to maturity was affected more by the environmental fluctuation. The regression coefficient ranged from 0.17 to 1.79, which showed instability for this trait in soybean during autumn (Table 8). Cultivars NARC-III and NARC-IV had mean values greater than grand mean and regression coefficients less than one indicating that these cultivars are stable under unfavorable environments. Cultivars Harper, NARC-V, SWAT-84 and William-82 had means greater than the grand mean and regression coefficients more than one; therefore, these cultivars are stable for late maturity. 58 During autumn, the mean of days to maturity ranged from 83 to 96 days with regression coefficients values form 0.17 to 1.79 along with varying degrees of deviation from regression (-l.54 to 1.96) which revealed that days to maturity is really difficult to predict. During spring (Table 9), the cultivars took more days to mature compared to the autumn season along with significant differences among genotypes- environments (location, year and their interaction). The mean values of the cultivars ranged from 118 to 127 days to maturity and regression coefficient value ranged from 0.97 to 1.06 with low to high deviations from regression (-1.44 to 2.71). For individual comparison of genotypes it was observed that Harper, NARC-III, NARC-IV, NARC-V, Swat-84 and William-82 had mean values more than the grand mean and regression coefficients close to one which shows the stability of the cultivars under all environments. Cultivars Century-84, FS-85 and Weber had regression coefficients close to one but they are associated with lower means than grand mean, therefore, these cultivars are stable for early maturity. This indicates that these cultivars showed less sensitivity to different environments (Harer and Shahi 1991). The stability parameters for all the sites (Table 10) indicate that Harper, NARC-III, NARC—IV, NARC-V Swat-84 and Williams-82 had above average (112) days to maturity. These cultivars possessed regression coefficients close to one which indicate that they are generally stable to all environments, whereas Century-84, FS-85 and Weber matured earlier than the average performance of all the cultivars overall (112) days but the regression coefficients of these cultivars is close to one which shows that these are stable. Maturity is one of the most important variables to be considered when developing new cultivars. Early maturing cultivars will fit into the multi—cropping systems of developing countries like Pakistan where water-holding capacity is low and short 59 duration cultivars are required to increase the cropping intensity. The short duration cultivars like Weber, Century-84 and FS-84 should be utilized in the breeding program in Pakistan for the development of cultivars suitable for the different cropping systems. 60 533 u w 52:30 n o . cosmos u ._ . amm> u > . _O>O_ £5205 6.0 a. 28556 : i¢®.N Or I I I I 0 X m X > :2: O - - - - o x O :OOO N - - - - O x > :3 FNNF F - - - - 6v 88% :FNO OF. :ON NO :3 OF > x ._ x 0 :3 ON :2. OF :sOF O o x O :3 OF :3 OF :OFN OF 0 x > :3: O :FNO O ..O.NON O 6:925 :OOOF O :32 v :ONO N ._ x > :36? O :OOOOF N :NNO F 2 8:80.. :3: N .5.va N :OON N E 59 3.? u. no 3.2 u ”a 33> a ma momaom jo 025.3. 2233 .32 BO OOOF .NOOF 9:3 .926 2O 9:9 5823 cc 3:29: 9 gap .2 20:82 v ocm 23> m .2958 cmogom m “—0 mocmtg ho m_m>_mc< N cam... 61 Table 8. Stability parameters [mean regression coefficient (b) and deviation from regression (S2 d)] of 9 soybean cultivars for days to maturity planted at 2 different locations during autumn 1992,1993 and 1994. Mean 92 92 96 96 96 96 96 83 95 Mean 94 St.error 0.32 CV 0.46 Tazble 9. Stability parameters [mean, regression coefficient (b) and deviation from regression (S2 d)] of 9 soybean cultivars for days to maturity planted at 2 different locations during spring1992,1993 and 1994. -84 Mean 125 Std.error 0.87 CV 1 .45 62 Table 10. Overall stability parameters [mean, regression coefficient (b) and deviation from regression (82d)] of 9 soybean cultivars, 4 locations, and 2 seasons for days to maturity during 1992, 1993 and 1994. Mean 1 0 1 1 14 1 4 14 14 115 1 113 Mean. 1 12 Std.error 0.85 CV 1.54 63 Plant height (cm): Plant height is affected by season as well as by year and location (Table 11). The pooled analysis of variance of cultivars, location, year and their interaction exhibit significant and non-significant results during spring, autumn and overall. Interactions year x cultivar and cultivar x location x year were found to be non-significant during autumn, spring and overall while year was significant during spring and non-significant in autumn. Overall analysis of variance for plant height at maturity also indicates non-significant as well as significant results. In overall analysis, year, year x cultivar and year x location x cultivar were non-significant. Similar results have been reported by Mebrahtu et al., (1991). Stability parameters estimated for plant height (Table 12) during autumn indicate that Harper, NARC-III, NARV-IV and NARC-V have regression coefficients < l which provide a measure of greater resistance to environmental change. Swat-84 and Williams showed regression coefficients > I with a mean greater than grand mean indicating increasing sensitivity to environmental change. NARC-III, NARC-IV and NARC-V had means above the grand mean and regression coefficients < 1 (Table 13). Ala (1990) stated that plant height was the most stable trait in maternal varieties while hybrids had higher coefficients of variance than standard cultivars. In the overall stability parameters (Table 14), the genotypes Harper, NARC-III, NARV-IV and Williams-82 were more stable under all environments demonstrated for their regression coefficients about 1 while Century-84, F S-85 and Weber were short stature than the grand mean and had regression coefficients < 1. Plant height is a very important parameter for any soybean cultivar, if a cultivator is too tall, it will lodge, thus reducing the yield. Therefore, a major objective 64 for the development of soybean cultivar is short stature. Weber, Century-84 and F S-84 have been found stable for short stature, therefore, these cultivars will be included in the breeding program for the incorporation of this variable in the development of a cultivar for potential regions in different cropping systems. 65 coOmmm u m 52:30 .I. o . 8:30.." ._ . bm0>fl > O .96. 8.0 a. 5.8:.ch : .95.. £59.05 8.0 a 28585 . OzONF OF - - - - o x O x > aim.” w I I I I 0 X W :O. FF N - - - - O x > :ONOFO F - - - - .O. OOOOOO OzN.F Ow OzOF NO Oz O.F OF > x O x 0 :ON ON .3 OF :NO O o x .. Oz FOO OF OzN.F OF OzOZ OF 0 x > ..O.NO O :ONN O :8? O 6. .9530 :3 O :FNF 4 OS N .. x > :OOO O :ONO N 83 F A... 8.80.. Oz OF N :NO N OzFoO N S .9; 253% “IE magi... Una 33> 1.... “IEI mom—40m .:o ozEnO szIB< .32 2O OOOF .NOOF OEOO .996 O5... O68 6532... O0 A83 599. Ema .8 Ocozmoo. v new 98> O .9928 $3.58 m .6 85:? O0 O_O>_mc< .FF 9an 66 Table 12. Stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] for plant height (cm) of 9 soybeans at 2 different locations during autumn 1992, 1993 and 1994 CULTIVARS Mean b S‘d CENTU RY-84 88 1.63 -1.48 FS-85 89 0.53 0.90 HARPER 94 -0.29 2.69 NARC-Ill 93 0.33 1.30 NARC-IV 93 0.26 -0. 19 NARC-V 94 0.11 0.66 SWAT-84 92 1 .68 1 .25 WEBER 81 1.92 -1.20 WILLIAM-82 92 2.83 -1.34 Mean 91 Std.error 0.34 CV 1.78 Table 13. Stability parameters [mean, regression coefficient (b) and deviation from regression (82d)] of 9 soybean cultivars for plant height (cm) of 3 locations during spring 1992, 1993 and 1994. AR Mean -84 .88 F 59 1.26 HARPER 61 1.98 ll 62 -1.4 -0.88 .02 T-84 61 0.49 WEBER 54 0.72 - . 5 Mean 60 Std.error 0.26 CV 2.47 67 Table 14. Overall stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars, 4 locations, and 2 seasons for plant height (cm) during 1992, 1993 and 1994. Mean -84 70 71 74 75 75 C- 75 A 74 65 LLIAM-82 73 Mean 72 St.error 0.68 CV 2.36 68 Pod height (cm): The pooled analysis of variance for distance of first pod from the ground is presented in Table 15. During autumn, only location showed non-significance while other sources like year, cultivar and their interaction exhibited a significant difference, demonstrating variability among cultivars and years. Analysis of variance for spring indicates significance for location and cultivar, but not for other factors. Analysis of variance for pod height on an overall basis revealed highly significant differences among the genotypes and environment, indicating the presence of variability among the cultivars as well as environment. Interactions year x location, season x cultivar, and year x season x cultivar were not significant. The estimates of stability parameters in autumn (Tablel6) showed that Harper, NARC-III, NARC-IV, NARC-V, Weber, and William-82 had regression coefficients > 1. Whereas Century-84, F S-85 and Swat-84 had regression coefficients < 1. All the cultivars had very low deviations from regression demonstratin that they are very stable for pod height. The stability parameter for pod height during spring (Table 17) revealed that Harper, NARC-III and NARC-IV had means greater than the grand mean, regression coefficients about 1 and very low deviation from regression. However, Century-84 and FS-85 had means less than the grand mean with regression coefficients 0.95 and 0.98 respectively (approximately one) showing average stability to all environments. Weber had a regression coefficient < 1 and a mean also less than the grand mean. Stability parameters simultaneously “considered in overall (Table 18) analysis for individual genotypic comparison revealed the genotypes, FS-85, Harper, Swat-84, and Williams-82 had means greater than the grand mean, with regression coefficients > 1, and 69 low deviation from regression indicating very good stability. NARC-III was greater than the grand mean with low regression coefficient (0.85) and deviation from regression (0.34). Whereas NARC-IV and NARC-V had means lower (5.4 and 5.5) respectively than grand mean with regression coefficients (0.89 and 0.83 respectively) and low deviations from regression which showed stability. Weber had a lower mean than the grand mean and regression coefficient < 1 with high deviation from regression whichshowed stability to changing environments. Pod height is very important variable in soybean yield. It depends upon variety, growth habit, maturity and the availability of planting space. If the pod height is optimum and within the reach of the combine, the yield will be increased, but if it is too close to the ground, the yield will be reduced. If the pod height is far from the ground, the yield will also be decreased. Therefore, such cultivars should be developed which have optimum pod height so that maximum yield can be achieved. 70 :OmNOm H m o 52:30 u o . c0383 .I. ._ . hm0> H > 0 2853.982 Oz .99 5529.. 8.0 a 28:23 . .96. £529.. SO 8 2856.9... OzO.F OF - - - - o x O x > OzO.F O - - - - o x O :NOF N - - - - O x > :ONNF F - - - - .O. 88$ .3 O4 OzOO NO :NO OF > x O x o :O .N O23 OF :FO O o x 4 :3 OF OzO.F OF :ON OF 0 x > :OO O :OO O :NOF O .9 .9530 Oz: O Ozao O :O N .. x > :2... O i. FOF N OzNN F 3. 8:83 :OOO N OzNO N :E N E .OO> 9...? u. an. 2:9, n. “a 3.9, u. “a momaoO ....o ozpam 2E3 .OOOF 2O OOOF .NOOF OEOO .996 2O 9:8 .cEausm .6 E3 Ego; con .2 Ocozmog v 96 O59. O .O..m>_O_:o cmmgoO m he 85...? .o 0.0>_mc< .mF mime 7l Table16. Stability parameters [mean, regression coefficient (b) and deviation from regression (82d)] for pod height (cm) of 9 soybeans at 2 different locations during autumn 1992, 1993 and 1994 .2 0.35 .42 0.75 0.32 Mean 7 Std.error 0.09 CV 1 0.52 Table 17. Stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] for pod height (cm) of 9 soybeans at 2 different locations during spring 1992, 1993 and 1994. 4 4 5 5 4 5 4 4 5 Mean 5 Std.error 0.06 CV 8.34 72 . .A’illi _ Table 18. Overall stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars, 4 locations, and 2 seasons for pod height (cm) during 1992, 1993 and 1994. Mean R -84 5 6 6 6 5 6 6 5 6 Mean 6 Std.error 0.08 CV 10. 34 73 Number of pods per plant: The number of pods per plant is a very important component of yield. The analysis of variance for number of pods per plant for autumn, spring and overall presented in Table 19 shows that during autumn, year, location and year x cultivar interaction were not significant while year x location, cultivar, location x cultivar, and cultivar x location x year were significant. In spring, there was no significant difference in location and location x cultivar interaction while the cultivars and other interactions showed significant differences; Highly significant differences among cultivars and their interactions on an overall basis suggested that the difference among the regression coefficients (b) of the cultivars on the environment means were real, suggesting that prediction across the environment is possible (Mebrehtu et al., 1991). The number of pods per plant during autumn (Table 20) indicates that NARC-IV had the highest number of pods per plant, a regression coefficient >1 with a low deviation from regression indicating that this genotype is sensitive to environmental fluctuations and greater specificity of adaptability to high number of pods per plant. Harper and Swat-84 had an average number of pods per plant less than the grand mean with high regression coefficients and low deviation from regression. NARC-III had the average number of pod per plant along with a low regression coefficient and deviation from regression indicating stability. During spring (Table 21), NARC-III and NARC-IV had above average means and their regression coefficients of about 1, which showed stability of these genotypes for number of pods per plant. Table 22 shows that NARC-III, NARC-IV and NARC-V had number of pods per plant above the grand mean and regression coefficients, 1.09,].17 and 1.05 respectively and 74 deviation from regression 1.50, 0.84 and -0.73 respectively, indicating good stability. FS- 85, Harper, Swat—84, Century-84, Weber and William-82 had mean values lower than the grand mean. Number of pods is a very important character which effects yield, thus, it is important to select the genetic material which is stable for the development of high yield potential cultivars suitable for different ecological zones under different cropping system in Pakistan. NARC-III, NARC-IV and NARC-V were found to be stable and had a high number of pods per plant. Therefore, these cultivars should be included in the breeding program for the development of desirable soybean cultivars for commercial cultivation in different regions of Pakistan. 75 5868 u m .9530 u c 50380..— H I. gm0> H > 6858.982 Oz .96. 5.5688 8.0 6 68586 . .96. 5.5688 SO 6 E8586 : OzO.F OF - - - - o x O x > .LO O - - - - OxO 1.5.OV N I I I I m X > satONN F I I I I AWV cowmmm :5 O4 .3 NO .3 OF C35 :NO 8 :3 OF :O.O O o x .. OzO.F OF OzN.F OF O20.F OF 9; :FON O :O. O O :ONO O .9 .9630 ....N.ON O :98 4 :52 N O x > OzOO O OzO.F N Ozao F 2 8583 :3: N .LNO N OzON N E 66> 9:9, n. Lo 629, u. no 3.9, 9 no lmomnoO 445550 02.5% 2.23:? .vmmw new mom? .39. 95:6 __m.m>o new 9.5% 65.336 50 5.63 .8 O08 ,6 69:2. .2 22.80. v new 9665 O .9928 F6238 O ho 8:965 56 O_O>_mc< .OF 633. 76 Table 20. Stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars for number of pods per plant planted at 2 different locations during autumn 1992, 1993 and 1994. Mean 33 Std.error 0.23 CV 3.18 Table 21. Stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars for number of pods per plant planted at 3 different locations during spring1992, 1993 and 1994. Mean 27 Std.error 0.21 Cv 6.20 77 Table 22. Overall stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars, 4 locations, and 2 seasons for number of pods per plant during 1992, 1993 and 1994. Mean -84 27 28 32 33 32 A 29 26 L 29 Mean 30 Std.error 0.20 CV 4.98 78 1”. -.-v—- ...v—. . Number of seeds per pod: The analysis of variance for number of seeds per pod during spring, autumn and overall are presented in Table 23 shows that there is no significant difference between cultivars and their interaction with location and year. It is evident from the analysis of variance that no differential response occurred among all the genotypes for number of seeds per pod. 100-Seed Weight (g): The analysis of variance for 100-seed weight for autumn, spring and overall presented in Table 24 shows highly significant differences among cultivars and interaction with year, location and year x cultivar. During autumn, cultivars were significantly different for 100-seed weight (g). Year x location, location x cultivar and year x location x cultivar also showed highly significant effects which demonstrated genetic differences among genotypes. This agrees with reports by Mebrahtu et al., 1991. During spring, the analysis of variance exhibited highly significant differences for all the interactions for 100-seed weight (gm) conducted at 3 locations and 3 years. Similar results were given by Mebrehtu et al. (1991) in soybean for 100 seed weight. The overall analysis of variance for lOO-seed weight (g) indicates that differences in cultivar, location and year and their interactions are highly significant. Stability parameters during autumn (Table 25), when simultaneously considered for the individual genotype revealed that NARC-III and NARC-IV had regression coefficients > 1 and low deviations from regression. The mean performance of these varieties was higher than the grand mean indicating stability to environmental fluctuations. 79 ‘5— '“ my . .l The stability parameters for 100-seed weight during spring are presented in Table 26. NARC - III and NARC — IV had means greater than the grand mean with regression coefficients higher than one and low deviations from regression. NARC—V had a mean greater than the grand mean with a regression coefficient less than one and low deviation from regression. On an overall basis for the stability parameters (Table 27) for 100-seed weight, Harper, NARC-III, NARC—IV and NARC—V showed the highest means and the greatest stability. Century—84, FS — 85, Swat — 84, Weber, Williams exhibited low lOO-seed weight and low regression coefficients. Developing cultivars with high stable 100-seed weight is a very important breeding objective. If there is any stress due to temperature or moisture, the lOO-seed weight will reduce the yield of the crop. Therefore, genetic material that has high stable seed weight should be included in the breeding program. NARC-IV and NARC-V have been found to have high 100-seed weight and good stability, therefore, these cultivars should be very useful in breeding for the development of high yielding soybean for cultivation in the different cropping systems in Pakistan. 80 tail-I. commem .I. m o .8530 u 0 . 50:600.. .I. ._ . 5N0> u > 0 6858.982 Oz OzOOO OF - - - - 0 x O x > szZ. O - - - - o x m w2omO N - - - - m x > OzOoO F - - - - .O. 88% O28... Ov O22... NO OzOOF OF > x O x 0 O22... ON OzOOO OF O28... O 0 x . OzFO.O OF OzOOO OF OzOOO OF 0 x > OzNOO O OzOF.F O OzNOO O .0. 9,530 OzO0.0 O OzOvO v OzOO.F N .. x > O200.0 O Oz F... N .28... F .... 8583 OzNOO N OOoN N OzONN N .>. 66> 3.9, n. .0 629, u. "a 629, n. .0 momaom .1256 02.5% 2.225... .82 .8 OOOF .NOOF 95.8 .996 O8 O88 .5533 .0 con .3 303 .0 .0955... .0. Ocosmoo. v .25 O69. O 902.30 509.3 m .0 0059.9 .0 O_O>_mc< .ON 036.. 81 0033 n m o .6280 u 0 . 00:000.. u 4 . 5mm> H > o mcmocEQOéoz wz .96. 5.589.. 8... 6 E8585 . .96. 5.589. FOO 6 E8586 .. itMN.m 0F I I I I o X m X > :NoO O - - - - 0 x O :88 N - - - - O x > .NOO F - - - - .O. 88% ..NO... O... :OOO NO ..ON.N OF > x 0 x 0 :8... ON ..O...O OF :OFO O 0 x . :OON OF ..OO.O OF OzOoF OF 0 x > rOvON O :OZN O ..N....OF O .0. 62.30 :ROO O :88 v :E. O N .. x > :NOO O :OO. FO N Oz OON F 3. 8580.. ..O0.0N N ..OO.NO N Oz Ono N E 66> 629, n. .0 98> . .0 9...; u. .0 00208 ....0 02.2% 2.20.5... .cEaSm .0 .3 £0.03 000909 .0. 9.03000. 0 000 200.. O 68F .8 OOOF .NOO. 9.8... .996 O8 95.8 .O.m>_._:0 000028 0 .0 0009.9 .0 99:05. .3. 0.00... 82 Table 25. Stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars for 100-seed weight (g) planted at 2 different locations during autumn 1992, 1993 and 1994. b .55 0.70 0.09 1 1 1.12 .76 0.25 1.79 1.43 Mean 18 Std.error 0.06 CV 2.25 Table 26. Stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars for 100-seed weight (g) planted at 3 different locations during spring 1992, 1993 and 1994. Mean -84 17 18 1 8 18 18 1 17 18 Mean 18 Std.error 0.04 CV 3.09 83 Table 27. Overall stability parameters [mean, regression coefficient (b) and deviation from regression (82d)] of 9 soybean cultivars, 4 locations, and 2 seasons for 100-seed weight (9) during 1992, 1993 and 1994. -84 Mean 1 8 Std.error 0.04 CV 2.79 84 0v 9 .I - U Oil content percentage: The pooled analysis of variance shown in Table 28 reveals a highly significant difference among cultivars and environments except in location during autumn and in season in the overall analysis, demonstrating the presence of variability among cultivars and environments. The mean values of oil percentage ranged from 20.8 to 21.3 (Table 29) during autumn. The coefficient of regression for the cultivars during autumn ranged from 0.66 to 1.32 with varying degrees of deviation from regression. The cultivars that had mean of oil content greater than the grand mean and a regression coefficient greater than one revealed stability under high oil content. Thus, Swat-84 had a mean greater than grand mean and regression coefficient > 1 showing stability for oil content. During spring (Table 30) NARC - III and NARC- V had high oil content and regression coefficients > 1 as well as low deviations from regression. Harper and Swat-84 showed high means and regression coefficients about 1 indicating good stability. Overall (Table 31) of all genotypes, Harper, NARC-III and Swat—84 produced oil content more than the average and had regression coefficients greater than one along with low deviation from regression indicating that these genotypes were stable for oil content. The oil percentage is a very important character for the development of soybean cultivars. The oil percentage is greatly affected by temperature during seed development (Howell and Cartter, 1953, 1958). Because of great importance of oil in Pakistan, cultivars should be developed that are stable and have high oil content. Harper, NARC-III, NARC-V and Swat-84 are stable and high in oil content, therefore, this genetic material should be used in the breeding program for the development of cultivars for commercial production. 85 ' " int-SM cowmmm H m e .9280 u 0 . COSNUOI— H J o hmm> H > 0 8858.982 O2 .96. 5.582.. SO 6 88586: 86 :ONO OF - - - - 0 x O x > {imm.m w I I I I 0 X m :OOON N - - - - O x > OzN0.0 F - - - - .O. 868O :NON O... :OON NO .83 OF > x .. x 0 :28 ON : FOO OF :O..N O 0 x .. :OOO O. :OO.F OF :8... OF 0 x > .LONF O :32 O :OOO O .0. .9530 :O... FF O :3O 4 :OOOF N .. x > :8. O :89 N OzNON F 3. 8.83 : .OOON N ..O..OO N :OOOvN N .>. 66> 629, u. .0 63.9, n. .0 9...; . .0 3208 ....$.m>0 02.5% 25.05.. .89 .08 OOOF .82 9:8 .696 .08 958 583.6 .0 .90. €6.80 .6 .0. 80.280. v new 999. O .9928 508.8 O .0 6029.9, .0 O_O.._Oc< .ON 032. quh 2 0‘9) C 1992, 1 : CULT|\ CENTL ,__————— I FS-85 " HA SWAT l WEBE . WlLLl‘- .l C") Mean Sid er CV Table 29. Stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars for oil content (%) planted at 2 different locations during autumn 1992, 1993 and 1994. Mean 20.9 Std.error 0.05 CV 1.56 Table 30. Stability parameters [mean, regression coefficient (b) and deviation from regression (82d)] of 9 soybean cultivars for oil content (%) planted at 3 different locations during spring 1992, 1993 and 1994. A Mean R -84 20. 20.8 R 21.1 1. 1.0 20. Mean 20.9 Std.error 0.02 CV 0.94 87 Table 31. Overall stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars, 4 locations, and 2 seasons for oil content (%) during 1992, 1993 and 1994. -84 BE Mean 20.9 Std.error 0.02 CV 1.23 88 5".- "I, Protei Protein content percentage: The analysis of variance for protein content for autumn, spring and overall basis for three years is presented in Table 32. No significant differences occurred during autumn or spring for either location or year x location. Location and season were also not significant for pooled data. When deviation from regression was considered as a measure of stability, protein percentage remained rather stable. For individual varietal comparisons, the linear regression could simply be stated as a measure of response of particular genotypes which is reflected by a number of environments. The deviation from regression was considered as a measure of stability of the variety with the lowest or non-significant standard deviation being the most stable. It is evident from the estimates of stability parameters (Table 31) that among all the genotypes tested in autumn, NARC-III, NARC-IV and NARC-V had high protein content and good stability. These three cultivars also had the highest protein content in the spring (Table 34). In Table 35, NARC-III, NARC-IV and NARC-V had the highest protein content overall. Soybean protein is also of high quality. This means that soybean is “sufficiently complete to sustain life for an extended period of time.” Therefore, it is very important for a developing country like Pakistan to develop cultivars which are high in quantity and quality of protein, because meat, eggs, and fish, other sources of protein, are very expensive and beyond the budget of ordinary persons. NARC-III, NARC-IV and NARC-V have high protein content and should be included in the breeding program in Pakistan. 89 :0O00m .I. m 02.30 H o . 02.003 u .— . .00> u > . 5.858.982 Oz .96. 85888 SO 6 E8580: i§N©.m QF I I I I 0 X m X > :8... O - - - - 0 x O :ONO N - - - - O x > Oz 80 . - - - - .O. 868O :FZV 9. :OOO NO :BO OF > x 0 x 0 :8... ON :NON OF :83 O 0 x 0 :8. OF :8. OF :85 OF 0 x > :88 O :OO...O O :88 O .0. .9530 :3. O 02 OO... O :OFNN N .. x > OzOON O Oz OO.F N Oz OOO F 3. 8.80.. :OO.OOF N .LOON. N :OEOF N .>. 66> 629, n. .0 8.9, n. .0 3.9, u. .0 00208 ....0 02.2% 2.20:5... .82 O8 OOOF .NOOF 85.8 .996 O8 9.86. 0.53.0 .0 .90. 2.0.000 50.0.0 .0. 2.05000. 0 000 900.. O .0.0>_._:0 0000.6»... 0 .0 00:0..0> .0 O_O>_0c< .NO 0.00... 90 Tazble 33. Stability parameters [mean, regression coefficient (b) and deviation from regression (32 d)] of 9 soybean cultivars for protein content (%) planted at 2 different locations during autumn 1992,1993 and 1994. Mean 40. 6 B. Std.error 0 34 "1 CV 0. 75 Tazble 34. Stability parameters [mean regression coefficient (b) and deviation from regression I (S2 d)] of 9 soybean cultivars for protein content (%) planted at 3 different locations during spring 1992 1993 and 1994. Mean R -84 40.2 40.4 .9 40. 40.9 40.5 40. Mean 40.5 Std.error 0.03 CV 0.54 91 Table 35. Overall stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars, 4 locations, and 2 seasons for protein content (%) during 1992, 1993 and 1994. d —84 .2 .24 -0. 40.3 1.13 -0.02 40.4 0.66 -0.03 40 .42 06 40.9 1.46 0.03 40.9 0.79 0.0 40.4 1.14 0.03 40.1 0.49 -0.01 40.6 1.67 -0.03 Mean 40.5 Std.error 0.02 CV 0.61 92 Yield (kg/ha): The pooled analysis of variance (Table 36) reveals highly significant differences among cultivars and environments (location and year) and their interactions for grain yield during autumn, spring and overall. Genetic differences among genotypes for their regression. on environmental index were non-significant. Diaz et al. (1990) Williams also found most cultivars stable for yield when sown in spring. The genetic differences among cultivars for regression coefficient on the environmental index were also observed by Eberhart and Russell (1966). However, they emphasized that both the regression coefficient and deviation from regression should be considered when analyzing the phenotypic stability of a particular genotype. Samuel et al. (1970) suggested that the linearity could simply be regarded as a measure of response of a particular genotype, which depends a number of environments, whereas the deviation from regression line was considered as a measure of stability. Thus, the genotype with the lowest deviation from regression being is the most stable. Table 37 shows that during autumn NARC-III, NARC-IV and NARC-V had high regression coefficients (1.95, 1.92 and 1.99 respectively) and means higher than the grand mean and also had low to medium deviations from regression, hence, they perform better in the high yielding environments. FS-85 and Swat-84 had regression coefficients of 1.07 and 1.29 respectively but had low yields. Harper had a yield lower than the grand mean and a low regression coefficient (0.047) and deviation from regression (-3 861 .61), indicating good stability in low yielding environments. The simultaneous consideration of three stability parameters (Table 38) during spring for the individual cultivars revealed that NARC-V and NARC-IV produced high 93 '.f' "..' ..‘ .3 iuig'kfi' r grain yield (2059 and 2049 kg/ha) and a regression coefficient around one (1.05 and 0.92) but very high deviation from regression indicating low stability. NARC-III produced a high grain yield (2049 kg/ha) and a lower regression coefficient (0.79) with low deviation from regression, indicating that this cultivar is especially good under unfavorable environments. Weber produced a grain yield (1853 kg/ha) and a regression coefficient close to one (1.10) but a high deviation from regression, indicating a high response to changes in the environments. FS-85, Harper and Swat-84 produced low grain yields and regression coefficients to above one with low deviation from regression (-2058.86, -1735.39 and — 1627.16 respectively), indicating that these cultivars may be considered good under favorable environments. The simultaneous consideration of stability parameters during overall analysis (Table 39) shows that NARC-IV and NARC-V produced higher grain yields (2248 and 2210 kg/ha respectively) and b =1.17 and 1.32 respectively, but high deviation from regression revealed that the cultivars performed better especially under favorable environments. NARC-III also had an above average grain yield and regression coefficient close to one (0.98), with a low degree of deviation from regression, indicating good stability. Yield has a highly significant interaction with genotypes and environments. High yielding cultivars need to be developed which have wide adaptability under the different cropping systems in Pakistan. NARC-III has wide adaptability; therefore, it should be used as genetic material commercial production in Pakistan. NARC-IV and NARC-V also have high yield and good stability. 94 nullify .I, ..i. 00003 n w . 0.02:5 u o . 00:000.. n ._ . umm> " > o .02.... £5290 80 00 0080090.. .92 £5290 50 0 00005090... .53 OF - - - - 0 x O x > :OFO O - - - - 0 x O :FONFN N - - - - O x > ..OOOO F - - - - FOO SO80 ..NOO Ow ..OvN NO :89 OF > x ._ x 0 :03 ON .OOO OF ..OOON O 0 x ._ ..OOO OF .03 OF :FOOF OF 0 x > 2.802 O ..NO.F.F O ..OOONN O 60 29030 ..FvOO O ..ONON v :SOF N 0 x > :NNNOv O :NOOFO N :FEO F :0 08003 :NOeO N ..OF. Fe N ..OONF N E 00> 83> u. “.0 3.2 0 ..E 3:; u. “.0 000000 ._._<0m_>o 02.000 255:2 .82 000 OOOF .NOOF 9.0% .0096 20 000% 000300 .0 059: 0.0; 00.. 0000002 0 000 0000.» O 0002030 000030 a 00 000000> 00 £02000. .00 0.03 95 Table 37. Stability parameters [mean, regression coefficient (b) and deviation from regression (Szd)] of 9 soybean cultivars for yield (kg/ha) planted at 2 different locations during autumn 1992, 1993 and 1994. d -84 .12 - .54 1.07 202.16 0.05 -3861.61 .95 - .92 .17 -1512.4 .85 1 -7 St.error 1 7.72 CV 4.93 Table 38. Stability parameters [mean, regression coefficient (b) and deviation from regression (82d)] of 9 soybean cultivars for yield (kg/ha) planted at 3 different locations during spring 1992, 1993 and 1994. -84 V A -84 BE Mean 1941 Std.error 18.92 CV 3.18 96 " “ 34.33 i Table 39. Overall stability parameters [mean, regression coefficient (b) and deviation from regression (82d)] of 9 soybean cultivars, 4 locations, and 2 seasons for yield (kg/ha) during 1992, 1993 and 1994. -84 - .82 -9154. -6752.21 159.2 7 .2 13906.56 -5487.07 13314.43 WI -10127.15 Mean 2061 Std.error 14.80 CV 5.78 97 SOCIAL CHALLENGES IN PLANT BREEDING TRADITIONAL PLANT BREEDING: In Pakistan, all the research institutes are operated by the government, therefore, government policies are followed. According to government policies, each program has its prescribed objectives and all the experiments are formulated according to those objectives. In order to achieve the objectives, efforts are being made to collect the germplasm and evaluate it. Afier evaluation, crosses of desirable characters are being made to get the desirable progeny. Genetic variability is created and promising lines are selected. Due to small quantity of availability of seed, only preliminary yield trials at the research stations are conducted. Then, National Coordinated Yield Trials of the promising lines at several locations are conducted. Then, demonstration trials are conducted on the farmer’s field trials. In the traditional breeding method, farmers are not involved in the selection and development of new cultivars. A participatory plant-breeding model was developed by means of which farmers; rural leaders, processors, and consumers are involved for the development of new cultivars. 99 Fig. 1. Outline of traditional plant breeding program in Pakistan TRADITIONAL PLANT BREEDING BREEDING OBJECTIVES (Government priorities) COLLECTION AND EVALUATION OF GERMPLASM GENETIC VARIABILITY/RECOMBINATION J I? x.“ ..‘" SELECTION OF PROMISING LINES TESTING OF NEW CULTIVARS RESEARCH STATION FARMERS FIELDS RELEASE OF NEW CULTIVARS 100 PARTICIPATORY PLANT BREEDING MODEL: In order for Pakistan to compete in global soybean production, an efficient and effective plant breeding model is needed. The new model will address the scientific, social, environmental and economic issues facing soybean production and utilization in Pakistan. The model will be developed for a public sector program, but many of the elements will also be relevant for a private sector breeding program. This model can also be used for other crops in other countries. The new model will outline seven activities that are important for a plant breeding program including (1) determining the breeding objectives, (2) collecting genetic diversity (3) generating variability/recombination, (4) selecting new cultivars, (5) testing new cultivars, (6) disseminating the new cultivars and (7) insuring sustainable funding for the program. Determining the breeding objectives is the first step in a seven to ten-year process of developing a new soybean cultivar. The breeder must interact with soybean farmers, consumers, processors, marketers and rural community leaders in identifying the important breeding priorities. The farmers will help to identify plant type, maturity, height, planting and harvest characteristics, biotic and abiotic stresses and other agronomic characteristics. Consumers will help to identify desirable color, taste, size, nutrition and texture traits. Marketers will identify the oil, protein and other value-added traits. The rural community leaders will help to identify environmental concerns as issues of equity and social justice. Will the cultivars be grown by the low resources farmers or high resources farmers? The rural community leaders will also look at potential value-added traits that can be bred into soybean in order to stimulate economic 101 growth in the rural communities. Speciality oils, proteins, and carbohydrates are potential value-added traits which could bring industries to the rural community. These industries would also bring much needed jobs to the rural communities. The next step is to prioritize the breeding objectives. Which characteristics are the most important and which will be addressed initially. Once the breeder does this, he/she must discuss prioritization of the breeding objectives with all stakeholders who helped develop the list of the breeding objectives? The breeder may need to adjust prioritization after reviewing it with the different groups. Getting feedback will be essential in developing trust with his/her clientale. Once the breeding objectives are determined, the next step is to determine the gennless that will be used in the crossing program. This will include soybean cultivars from breeding program as well as cultivars from other breeding programs in Pakistan and other countries. Wild species of soybean may also be needed for the integration of desirable characteristics. Genetic material from international gene banks will also be needed. AVRDC, IPGRI and USDA have extensive soybean collections from which to access material. If the breeding program uses genetic engineering, then several other questions must be addressed. The first is whether the farmers will grow and /or marketers will buy the genetically modified soybean. The next is where will be got the desired genes? What promoters will be used? What terminators will be used? Do you have freedom to operate? Are there intellectual property issues that must be added? What bio-safety procedures will need to be followed? 102 Step three is to generate the new segregating populations. Traditional hybridization, mutagenesis or genetic engineering can be used. Some breeding programs will perform a combination of the above. The selection process is a very complex and important step. It is important to incorporate as much science as possible to help insure success. At the same time, any process that involves multi-stakeholders in the development of new cultivars will help increase the probability of success. Screening large numbers efficiently is crucial in the development of new cultivars. The breeder must identify the different screening procedures that will be used and will need to screen large segregating populations for resistance to insects and diseases as well other traits. Coordination with plant pathologists, entomologists, chemists and other related scientists is important. Screening for biotic and abiotic stresses can be done in the field as well as the laboratory or green house. Farmers should be involved once some of the initial screening for the biotic and abiotic stresses is completed. Farmer selection should occur once the segregating populations have more or less stabilized (F4 generation). Farmers can add valuable inputs into selecting material that will be entered into preliminary yield trials. There is considerable literature on ways of involving farmers in selection. Preliminary yield tests can then be tested at the experiment station. Again, farmers can be involved in rating the promising lines. However, breeders and the farmers must determine the cost and benefit of this process. Advanced yield trials should be carried out on the experiment station as well as on several farmers’ fields. Multiple locations will help to identify the stability of the 103 ' . v.30 IrrumJ important traits, i.e. yield, oil content, pest resistance etc. At least two years of advanced yield data is suggested before releasing a new cultivar. Field days are important. In this way farmers have the opportunity to visit the fields of new culitvars and also have a chance to see their performance. During the field day, dissemination of the latest production technology is also displayed and the farmers become oriented to this technology. Farmers also become aware of new cultural practices, insecticides, herbicides and other chemicals needed to achieve maximum yield. Extension workers also need to be involved in the cultivar testing program. They can play an important role in the introduction of new soybean cultivars, especially in fallow areas by helping to disseminate production technology to farmers with the colloboration of the research workers. Also, finding ways to involve private seed companies is important. Soybean has a lot of potential and can fit in the different cropping systems under different agrological zones of Pakistan; therefore, involvement of private companies of seed will be very helpful to increase the area under soybean. The public sector breeder should find ways to enhance the involvement of the private sector. This could be in providing enhanced germplasm for private breeding programs as well as encouraging the private sector to market cultivars. Licensing new cultivars to private companies can be very effective in getting new cultivars to the farmers. An effective breeding program must find ways to increase the necessary funding for a successful program. Working with the farmers, industry, marketers and rural leaders will provide opportunities to get fimding for the breeding program. This group will help to find funds, both public and private, for the program. In addition to the public 104 v .‘ ~&... I- m... T sector funds, financial support form farmers groups, industries, foundations and seed companies are possible. Involving these people will also help keep a long-term vision for the breeding program. This group is very interested in finding new ways to participate in the global economy. 105 Fig. 2. Outline of Participatory Plant Breeding Model for a soybean breeding program in Pakistan. PARTICIPATORY PLANT BREEDING MODEL CONSUMERS PROCESSOR FARMERS BREEDING OBJECTIVES l RURAL LEADERS COLLECTION & EVALUATION OF GERMPLASM PROCESSORS CONSUMERS 7 SELECTION OF PROMISIN GENETIC VARIABILITY/RECOMBINATIO N (traditional /biotechnology) LINES RURAL LEADERS FARMERS TESTING OF NEW CULTIVARS l EXPERIMENT STATION FARMERS’FIELD MULTILOCATION TRIALS FARMERS, RURAL COMMUNITY LEADERS AND CONSUMERS SOYBEAN BREEDERS, — RESEARCHERS AND EXTENSION WORKERS PROCESSORS AND MARKETERS l DISSIMINATION OF NEW CULTIVARS 106 SUMMARY CHAPTER-V SUMMARY Keeping in view the significance of soybean as an oilseed crop in the country and effect of genotype x environmenr (G x E) interaction on variety performance. This present study was conducted to identify high yielding, stable genotypes along with other agronomic traits of nine genetically different genotypes of soybean (Glycine max). The experiments were conducted in two seasons during spring and autumn at four locations for three years 1992-94. The experiments were planted at all locations during both seasons of each year in a randomized complete block design with four replications. Plot size was four rows with 5 meters long and row spacing in autumn was 45 cm and 30 cm in spring. Data were recorded for days to flowering, days to maturity, plant height, pod height, pods per plant, seeds per pod, 100 seed weight, oil percent, protein percent and yield. The pooled analysis of variance for days to flowering and maturity for autumn, spring and overall basis revealed significant differences among genotypes, environment and their interaction. Days to flowering and maturity were more affected by seasons as during autumn the crop took less days as compared to spring. Weber was found to be early in flowering and maturity and stable in all the environments. Therefore, Weber can be used as genetic material for the development of early varieties. Highly significant differences in pooled analysis for spring, autumn and overall basis were found for plant height. ‘NARC-Ill’, ‘NARC-IV‘ and ‘NARC-V’ attained the maximum plant height during autumn and spring as well as overall. Therefore, these lines can 108 C01 be used for the breeding program. Highly significant differences among genotypes, environment and their interaction were observed in pooled analysis of variance for spring, autumn and overall, indicating the presence of variability among the genotype as well as environment and their regression on environmental index for first mature pod height and pods per plant. During autumn, ‘Swat-84’ exhibited maximum pod height while ‘NARC-IV’ exhibited maximum pods per plant. During spring, ‘NARC-lll’ exhibited maximum pod height where as ‘NARC-V’ showed the highest number of pods per plant. This genetic material can be used in the breeding program for the development of cultivars with these characteristics. Non-significant variability was observed for number of seeds per pod. Pooled analysis of variance of 100-seed weight for spring, autumn and overall showed highly significant difference among genotypes, environment and their interaction, indicating the presence of both predictable and non-predictable components of G x E interaction. During both seasons as well as on overall, ‘NARC-lll’, ‘NARC-IV' and ‘NARC-V’ attained the maximum 100-seed weight and these lines can be useful for the development of cultivars having maximum grain weight. The pooled analysis for autumn, spring and overall revealed highly significant difference among genotypes, environments and their interaction for oil and protein content percentages. Overall, Swat-84 showed the maximum oil percentage while NARC-V had the highest protein content. These lines can be 109 useful for the development of cultivars with high oil and protein but in separate breeding programs. There were highly significant differences among the genotypes and environments for grain yield (kg/ha) during spring, autumn seasons and overall. ‘NARC-lll’, ‘NARC-IV’ and ‘NARC-V’ had the highest yield and showed good stability for the favorable environments. A participatory Plant breeding model was developed which addressed the scientific, social, environmental and economic issues facing soybean production and utilization in Pakistan. The model consists of seven steps (1) determining the breeding objectives (2) collecting genetic diversity (3) generating variability/recombinations (4) selecting new cultivars (5) testing new cultivars (6) disseminating the new cultivars and (7) insuring sustainable funding for the program needed for the development of plant breeding program by means of which cultivars can be developed for different ecological zones. This model will be useful not only to increase soybean production in Pakistan but also for the development of other commodities. 110 ‘k l -I ,. LITERATURE CITED LITERATURE CITED Abou—El—Fittouh, Rawlings, J0. and Miller. PA. 1969. Genotype by environment interactions in cotton-their nature and related environmental variables. Crop Sci 9 (3): 377-381. Adair, CR. and Jones, J.W. 1946. Effects of environments on the characteristics of plant surviving in bulk populations of rice. J. Ames. Soc. Agron. 38:708-716. Akhtar, H.R., Byerlee, D. Qayyum, A., Majid, A. and Hobbs PR. 1986. Wheat in the cotton. Wheat F arming System of the Punjab. P. l- 4. Alanis, L. Bucto. 1966. Environmental and genotype-environmental components of variability. Inbred lines. Heredity. 21:387-397. Ala-VS. 1990. Radiative in the main characters in F3 hybrids obtained by crossing mutants with wild soybean. Pla-Bre. Abs. 1993. 063-03977. Allard, R.W., and Bradshaw, AD. 1964. Implication of genotype x environmental interactions in applied plant breeding. Crop Sci. 4 (6): 503-508. Association of Official Seed analysis. 1991. Rules for testing seeds. Journal of Seed Technology 12(3):1-109. Baihaki, A., Stucker, RE. and Lambert, J .W. 1976. Association of genotype x environment interactions with performance level of soybean lines in preliminary yield tests. Crop Sci. 16(5): 718-721. Baker, R. J. 1969. Genotype — Environment interaction in yield of wheat. Can. J. Plant Sci. 49:743-751. Borojevic, S., and Williams, WA. 1982. Genotype x Environment interaction for leaf area parameters and yield components and their effects on wheat yields. Crop Sci. 22(5): 1020 1025. Breese, EL. 1969. The measurement and significance of genotype environment interactions in grasses. Heredity-24: 27-44. Brennan, RS. and Byth, DE. 1979. Genotype x environmental interactions for wheat yields and selection for widely adapted wheat genotypes. Aust. J. Agri. Res. 30: 221—232. Byerlee, D., Sheikh, A.D., Aslam, M. and Hobbs, PR. 1984. Wheat in the Rice—Based Farming System of the Punjab. Implementation for Research and Extension. P. 1- 49. 112 Carver, B.F., Burton, J .W., Carter, Te. Jr. and Wilson, RF. 1986. Response to environmental variation of soybean lines selected for altered unsaturated fatty acid composition. Crop Sci. 26(6): 1176-1181. Chan, K.L., Lee, R-s. and Thseng, F-s. 1986. Genetic variation of soybean hybrid population by natural selection in spring crop season at four locations 1. Frequency distributions for certain agronomic characters in F2 generation. J. of Agri. Res. of China, 35(3): 263-279. Cianzio, S.R de. Green, D.E., Chang, C. -S. and Shibles, RM. 1991. Developmental periods in soybean: Photoperiod sensitive x insensitive crosses evaluated at diverse latitude. Crop Sci.31 (1): 8-13. Clark, P. and Snyder, H. 1989. Effect of location and growing season on oil content of TEL soybean cultivars. Arkansas-Farm Research. 38(4): 2-3. ' Comstock, RE, and Robinson, HF. 1952. Genetic parameters, their estimation, and significance. Proc. 6th Int. Grass. Congr. p 284-291. F -— Comstock, RE, and M01], RH. 1963. Genotype x Environment interactions. Symposium on stabilitical genetics and plant breeding. NAS-CRC. Pub. 982, P. 164-196. Deeprose, D. 1995. The team coach. New York: AMACOM. Diaz-Carrasco, H., Gonzalez—Mauri, H. and Relazguez, O. (1990). Adaptability and stability of yield in spring soybeans. Ciencias-de-la-Agricultural. 39:173-175. Eberhart, SA. and Russell, WA. 1966. Stability parameters for comparing varieties. Crop Sci. 6 (1): 36-40. F.A.O. 1999. Production yearboolg F.A.O. 1999. Food balance sheet (http://appsfaoorg/lim500/nph- wra. . . lanceSheet&Domain-FoodBalanceSheet). Ferriss, RS. and Baker, J .M. 1990. Relationship between soybean seed quality and performance in soil. Seed Sci. and Tech. 18:51-73. Finlay, K.W., and Wilkinson, G.N. 1963. The analysis of adaptation in a plant breeding programme. Aust. J. Agric. Res.14:742-754. Fisher, K. 1993. Leading self-directed work teams. New York: McGraw-Hill. Goth, K. 1955. Genetic analysis of varietal differentiation in cereals. 1. Statistical differences found among local strains of a barley variety, “Hosofara No. 2”. Japan Genet. 30: 95-106. 113 Grafius, J .E. 1956. Components of yield in oats. A Geometrical. Interpretation. Agron. J. 48 (9): 419-423. Green, D.E., Shibles RM. and B.J. Moraghan. 1972. Breeding for improved yield via morphological characteristics. Proc.Soybean Seed Res. Conf. 2: 43-46. Hanson, W.D. 1970. Genotypic stability. Theor. Appl. Genet. 40: 226-231. Harer, RN. and Shah, BS. 1991. Stability for seed yield in some soybean varieties. Legume Res. 14(4): 195-196. Hildebrand, D.F., Hymovvitz, T. and Weber, El. 1988. Soybean oil stability "effect of genotypes environment and heat denaturation of seed components". Soybean Genetics News Letter 15: 96-102. Hobbs, P. R., Khan, B. R., Nasir, M. Nasir, and Munir, M.Munir. 1983. Yield assessment of wheat in barrani (rainfed) area of Rawalpindi District. NARC, Wheat Program, Islamabad. Hobbs, P.R., Khan, B.R., Munir, M, Razzaq, A. and Hashmi, NI. 1984. Agronomic results for wheat in rainfed area of Northern Punjab. NARC, Islamabad. Homer, T. W. and Frey, K.J. 1957. Methods for determining material areas for oat varietal recommendations. Agron. J. 49(6): 313 — 315. lman, AG. and Allard, R.W. 1965. Population studies in predominantly self-pollinated species. VI. Genetic variability between and within natural population of wild oats from different habitats in Califomia. Genetics. 51: 49-62. Ivory, D.A., Karwmeechai, S., Delacy, I.H. and Basford, KB. 1991. Analysis of the environmental component of genotype x environment evaluation in crop adaptation. Field Crops Res. 28: 71-84. Jaap Hardon. 1995. The Global Context: Breeding and crop genetic diversity. Participatory Plant Breeding. P 1-2. July 26-29. Jensen and Neal. F. 1952. Intra varietal diversification in oat breeding. Agron. J. 44: 30-34. Johnson, V. A., Shafer, L. and Schmidt, J. W. 1968. Regression analysis of general adaptation in hard red winter wheat (Triticum aestivmn L.). Crop Sci. 8(2): 187 - 191. Jowett, D. 1972. Yield stability parameters for sorghum in East Afiica. Crop Sci.12 (3): 3 14-3 17. 114 Konwar, BK. and Talukdar, P. 1986. Stability analysis of yield and its components in soybean. Crop Improvement 13(1): 172-175. Kroonenberg, P.M., and Basford, KB 1989. An investigation of Multi- attribute genotype response across environments using three-mode principal component analysis. Euphytica. 44: 109-123. Lal, V.S., and Haque, M.F. 1971. Path analysis of yield components in soybean. Indian J. Genet. Plant Breed. 31: 357-362. Lawn, R.J. and Imrie, BC. 1991. Imrie, Crop improvement for tropical and subtropical Australia, Designing Plants for different climates. Field Crops-Research. 26: 113- 139. Leholm, A., Larry, H., Murari, S., G.Ian. and Fred, P. 1999. Area of Expertise Teams: The Michigan Approach to Applied Research and Extension. 37(3): 1-9. Luthra, O. P., Singh, R. K. and Kahar, S. N. 1974. Comparison of different stability models in wheat. Theor & Appl. Gen. 45: 143 — 149. Mariani, B.M. and Novaro Manmana, P. 1986. Combination of results from several data sets: some possibilities for a better understanding of the genotype-environment interaction. Genet. Agr. 40: 83-96. Mayers, J.D., Lawn, R.J. and Byth, DE. 1991. Adaptation of soybean (Glycine max L. Merrill to the dry season of the tropics II. Genotypic and Environmental effect on biomass and seed yield. Aust. J. Agric. Res. 42(3): 517-530. Mebrahtu, T., Mohammad, A., and Mersie, W. 1991. Green pod yield and architectural traits of selected vegetable soybean genotypes. J. Prod. Agri. 4(3): 395-3 99. Molari, P., M. Lucchin, and Parrini, P. 1987. Effectiveness of early generation selection on SSD soybean (G.max) populations. Genetica-Agraria 41(3): 306. Naidu, N.V., Satyanarayana, A. and Habibuddin, K. 1991. Stability analysis for seed yield and its components in mungbean [Vigna Radiata (L.) Wilczek]. Legume Res. 14(4): 169-174. Odendaal, T.E.M. and Deventer, C.S. van. 1987. Genotype- Environment interaction in soybean. South Afiican J. of Plant & Soil 4(2): 65-69. Orsubum. J., Moran, L. Musslewhite, E. and Zenger, J .H. 1990. Self-directed work teams: The new American challenges. Burr Ridge. IL. Irwin. 115 Pakistan Economic Survey. 1998-99. Area and Production under Important Crops. P. 23. Pakistan Economic Survey. 1998-99. Sectoral Share in GDP. P. 4-5. Pathak, MN. and Nema. DP. 1984. Genotype-environment interaction for growth characteristics and yield in soybean. Progress in Ecology Vol. 10 Edited by Verma G.P. 143-163. New Dehli, India, Today + Tomorrow. Patil, HS. and Narkhede, B.N. 1989. Phenotypic stability of yield, seeds/pod and 100 seed weight in blackgrarn (Gina mungo. Hepper). Legume Research. 12(4): 189-192. Patil, V.P., Raut, V.M., Halvankar, GB. and Taware, SP. 1989. Stability analysis for seed yield and its components in soybean. Proc. World Soybean Res. Conference IV, P. 1042-1047. Perkins, J. M., and Jinks, J. L. 1968. Environmental and genotype environmental components of variability. 111 Multiple lines and .crosses. Heredity 23: 339 — 356. Peter Golbitz. 1999. Sova & Oilseed BLUEBOOK. P. 330 Pfeiffer, T.W. 2000. Selecting Soybean for Adaptation to Double Cropping on the Basis of Full Season Plant height. Crop Sci. 40(2): 387-390. Pfieffer, W.H., and Braun, H.J. 1989. Yield stability in bread wheat. In.J.R. Henderson and PB. Hazell (ed.) Variability in grain yield. Implications for agricultural research and policy developing countries. The John Hopkins University Press, Baltimore. Plaisted, R. L. and Peterson, L. C. 1959. A technique for evaluating the ability of selections to yield consistently in different locations or seasons. American Potato Journal, 3 6: 3 8 1 -3 85. Pookpakdi, A. and Siriprasert R. 1986. Days to flower, day to complete flowering and flowering period of Thai soybean varieties as affected by extended photoperiods. P1. Breed. Abst. 058-08068. Porter, P. M., Lauer, J. G., Lueschen, W. E., Ford, J. H., Hoverstad, T. R., Oplinger, E. S. and R. K. Crookston. 1997. Environment Affects the corn and Soybean rotation Effect. Agron. J. 89(3): 442 — 448. Qari, M.S., Chaudhry, M.H., Khan, N.I., Hussain, B. and R. Din. 1990. Phenotypic stability in barley. J. Agric. Res. 28(3): 227-231. Quick, T. 1992. Successful team building. New York: AMACOM. 116 Reicosky, D.A., Orf, J .H. and C. Poneleit. 1982. Soybean germplasm evaluation for length of seed filling period. Crop Sci. 22(2): 319-322. Riaz A.K., Hashmi, N.I., Nasir. M.,Munir,M.and P.R.Hobbs. 1983. Farming System Research. An exploratory survey of rainfed agriculture in Northern Punjab. PARC Staff, Paper AE-103. Rosielle, AA, and Hamblin, J. 1981. Theoretical aspects of selection for yield in stress and non-stress environments. Crop Sci. 21(6): 943-946. Saeed, M. and Francis, CA. 1984. Association of weather variables with genotype x environment interaction in grain sorghum. Crop Sci. 24 (17): 13-16. Saeed, M., Francis, CA. and Rajewski, J .F. 1984. Maturity effects on genotypes x environment interaction in grain sorghum. Agron.J.76 (1): 55-58. Saini, M. L., Johrar, B. B. and Solanki. R. K. 1977. Genotype x Environment interaction for seed yields in cluster bean. Indian J. Agri. Sci. 47(7): 345 — 347. Samuel, C.J.A., Hill, J ., Breese, FL. and Davis, A. 1970. Assessing and predicting environmental response in Lolium perenne L. J. Agric. Sci., Camb. 75: 1-9. Schnebly, SR. and Fehr, W.R.1993. Effect of years and planting dates on fatty acid composition of soybean genotypes. Crop Science 33(4): 716-719. Schoorel, AF. 1957. The use of soil tests in seed testing. Proceeding of the International Seed Testing Association. 22: 287-302. Sen, S.N., Mukherjee. 1986. Suitability of soybean cultivars to different seasons. Presented in XVII All India Soybean Res. Workshop. MACS, Pune 22-25. Sharma, SK. and Mehta H. 1988. Effect of cropping system on genetic variability and component analysis in soybean. Field Crop Res. 19(2): 103-111. Singh, 0. and Chaudhry, B.D. 1985. Stability analysis for protein and oil content in soybean. J. of Oilseeds Res. 2(1): 57-61. Specht, J .E. and Williams, J .H. 1978. Testing soybean for heat and drought tolerants. Proc.Soybean Seed Res. Conf. 8:17-32. Steele, R.G.D. and Torrie, J.H. 1960. Principal and procedures of statistics. McGraw Hill Book Co. Inc. New York. Suvedi, M. 1996. Farmers’ perspectives on Michigan State University Extension: Smnmary report. East Lansing: Michigan State University, Extension Program Support Systems. 117 Verma, O.P., Singh, RS, and. Vishwakarma, N., Vishwakanna. 1987. Phenotypic stability of seed yields in Barley. Rachis. 6(1): 24-26. Wang, H. and Murphy, P. A. 1994. Isoflavon composition of American and Japanese soybeans in Iowa: Effects of variety , crop year, and location. J. Agri. Food Chem. 42:1674-1677. Weiltzien-RE, Whitmaker ML, Rattunde HF W, Dharnotharan M, Aners MM. 1998. Participatory approaches in pearl millet breeding. In: Witcombe JR; Virk DS; Farrington J, eds. Choice of seed. Making the most of new varieties for small farmers. Oxford and [EH Publ. Co., New Delhi, Calcutta p 143-170. Wellins, R.S., Bhyam, W.C., and Wilson, J.M. 1991. Empowered: Creating self-directed work groups that improve quality, productivity, and participation. San Francisco: Josey —Bass. 1999. Double cropping wheat and soybeans. P. 143-156. In L.G. Heatherly and HF. Hodges (ed.). Soybean production in the mid south. CRC Press. Boca Raton, FL. Whitehead, W.F. and Allen, FL. 1990. High vs low stress yields test environments for selecting superior soybean lines. Crop Sci. 30(4): 912-918. Witcombe JR, Joshi A., Joshi, KD. and B. Sthapit. 1996. Farmer participatory crop improvement 1. Methods for varietal selection and breeding and their impact on biodiversity. Expt. Agri. 32: 445-460. Woodstock, J .W. 1973. Physiological and biochemical tests for seed vigor. Seed Science and Technology. 1:127-157. Yadar, LS. and Kurnar, D. 1983. Stability for seed boldness in black gram. Madras Agric. J. 70:193-194. Yates, F ., and Cochran, WC. 1938. The analysis of groups of experiments. Journal of Agricultural Science, 28: 556-580. Zhu-J.C. 1992. Study on the heritability, genetic advance and correlation of primary agronomic traits of spring soybean varieties sown in spring and autumn. Soybean Sci.1l (4): 322-328. 118 APPENDIX Table 1A. Structural changes in area of oilseed crops in Pakistan 1970- 75 to 1998-99. Year Total oilseed Total cropped Oilseeds area area (000,ha) area (000,ha) (%) 1970-75 603 17.16 3.5 1975-76 562 18.02 3.1 1976-77 619 18.21 3.4 1977-78 525 18.49 2.8 1978-79 578 19.30 3.0 1979-80 534 19.22 2.8 1980-81 552 19.33 2.9 1981-82 541 19.78 2.7 1982-83 540 20.13 2.7 1983-84 463 20.06 2.3 1984-85 469 19.92 2.4 1985-86 471 20.28 2.3 1986-87 441 20.90 2.1 1987-88 400 19.52 2.1 1988-89 459 21.82 2.1 1989-90 453 21.30 2.1 1990-91 497 21.82 2.3 1991-92 496 21.72 2.3 1992-93 473 22.44 2.1 1993=94 512 21.87 2.3 1994-95 499 22.14 2.3 1995-96 569 22.59 2.5 1996-97 614 22.93 2.7 1997-98 664 23.04 2.9 1998-99 615 23.04 2.7 Source: Pakistan Economic Survey. 1998-99. Agricultural Statistics of Pakistan. 1970-91. FAO. STAT Database results. 1991-99 (www.FA0.0rg) 120 Table 2A. Share of domestic production and import in the total availability of edible oil in Pakistan 1970-75 TO 1998-99. Year Dom.Prd T. Req. P. C. Import Import . (000,t) (000,t) Cm.kglyear (000,t) (%) 1970-75 201 307 GT5 127 38 1975—76 218 435 8.0 217 50 1976-77 228 496 9.2 268 54 1977-78 232 536 8.7 304 57 1978-79 261 602 9.0 341 57 1979—80 290 650 9.7 360 55 1980-81 271 698 10.0 427 61 1981-82 280 794 10.3 514 65 1982-83 219 870 1 1.1 651 75 1983-84 305 920 1 1.6 615 67 1984-85 349 1038 10.8 689 66 1985-86 325 1 159 10.6 834 72 1986-87 462 1433 10.9 971 68 1987-88 482 1385 1 1.5 903 65 1988-89 478 1294 13.8 1016 79 1989-90 458 1359 12.6 901 66 1990-91 497 1449 13.0 962 66 1991 -92 598 1676 13.0 1078 64 1992-93 473 1791 14.8 1318 74 1993-94 393 1846 15.3 1453 79 1994-95 450 1885 13.8 1435 76 1995-96 525 1812 15.7 1287 71 1996-97 568 1831 14.9 1263 69 1997-98 514 1844 14.0 1330 72 1998-99 532 1868 14.6 1336 72 Source: FAO.Food Balance Sheet 1971-99. www.FA0.0rg Dom.Prd. Domestic Production Total Req. Total requirement P. C. Con. Per capita consumption 121 Table 3A. Conventional and non-conventional source of edible oil in Pakistan during 1999. Conventional Prod. (000, Oil (°/o)* Total oil (000, Total (%) tons) tons) Cottonseed 3824.00 13 497.12 71 .54 Rape-Mustard 282.00 35 98.7 14.20 Goundnut 55 31 17.05 2.45 Sesame 35 44 15.4 2.22 Non-conventional Sunflower 195 33 64.35 9.26 Soybean“ 10 19 1.9 0.27 Safflower 1.2 33 0.4 0.06 Source: . FAO. Food Balance Sheet. 1999. *Expeller "Solvent extraction Table 4A. Fatty acid composition of different vegetable oils. Crop Un-S. F. A. (%) S. F. A. (%) Other S. F. A. (%) Linoleic Oleic Stearic Palmitic Safflower 78 15 2 5 1 Sunflower 68 21 5 6 1 1 Soybean 55 21 6 9 6 Corn 55 30 3 8 3 Cotton 51 22 2 2 1 Peanut 31 50 6 8 - Sesame 55 21 6 9 1 Palm 8 42 5 41 - Coconut* 2 7 2 9 - Butter 3 35 12 27 3 Ra.& Mustard 14 18 1.5 3.5 63* * Lauric acid 48%, other saturated acids 32% * * Erucic acid 40% Source: 122 Thieme, JG. (1968). Coconut oil processing. FAO, Rome Food. The Year Book of Agriculture. 1959, USDA. S. F. A. Saturated fatty acid Ra. & Mustard Rapeseed and Mustard Table 5A. Existing cropping in different areas of Pakistan. Area Rotation Rice area Rice Wheat Rice Rice Fodder Rice Rice Fallow Rice Rice Matri Rice (June-Nove) (Oct-May) (June-Nove) Cotton Area Cotton Wheat Cotton Cotton Fallow Cotton Cotton Fodder Cotton Cotton Rapeseed and Cotton Mustard (June-Dec) (Oct-May) (June-Dec) Rainfed Area Wheat Fallow Wheat/Ra.and Mustard (Oct-May) (Oct-May) Sorghum Fallow Wheat (July-Oct) (Oct-May) Source: 0 Wheat in the cotton-wheat farming system by Akhter, H.R., D. Byerlee, A. Qayyum, A. Majid and PR Hobbs. 1986. P. 1-64. . Agronomic results for wheat in rainfed area of Northern Punjab. NARC, Islamabad. Table 6A. Proposed cropping system for soybean in different areas of Pakistan. Area Rotation Rice fallow area Rice Soybean Rice (June-November) (February-May) (June-November) Cotton fallow area Cotton Soybean Cotton (June-December) (February-May) (June-December) Rainfed area Wheat Soybean Wheat (October-May) (July-October) (October-May) 123 Table 7A. Agro-meteorological data at NARC, Islamabad during 1992, 1993 and 1994. Latitude 33° 40' N Longitude 73° 10' E Average Temperature (°C) Rainfall (mm) Month 1992 1993 1994 1992 1 993 1 994 Max Min Max Min Max Min January 15.4 5.5 15.2 2.9 17.1 5.0 119.1 28.3 35.1 February 17.9 5.5 21.2 7.0 17.1 6.0 83.5 46.9 47.6 March 21.7 9.2 21.1 8.3 25.6 10.8 108.7 144.7 24.8 April 27.1 13.5 29.2 14.2 28.0 12.6 46.9 27.8 70.0 May 31.9 17.2 36.4 20.3 36.0 19.5 52.5 23.6 44.0 June 37.1 22.2 37.7 23.1 40.2 23.4 23.6 83.2 69.1 July 33.4 23.9 33.1 23.7 33.1 24.1 155.6 262.5 535.1 August 32.0 24.3 33.9 23.4 31.8 23.4 182.8 224.9 578.5 September 30.4 20.5 30.6 20. 5 32.1 18.1 358.4 228.8 76.6 October 28.5 13.9 29.8 12.5 28.7 11.8 12.8 0.0 44.5 November 22.9 8.3 24.8 8.7 25.2 7.6 35.0 8.7 2.2 December 19.1 6.1 21.3 3.6 17.5 4.8 8.0 0.0 126.0 Source: - Water Resources Research Institute, NARC, Islamabad. 124 Table 8A. Agro-meteorological data at Fatehjang during, 1992, 1993 and 1994. Latitude 33° 34' N Longitude 720 39' E Average Temperature (°C) Rainfall (mm) Month 1 992 1993 1 994 1 992 1993 1994 Max Min Max Min Max Min January 15.48 6.67 14.19 3.77 15.7 6.3 98.25 23.87 30.40 February 16.17 6.41 20.25 7.75 17.3 6.9 94.97 26.89 45.5 March 21.19 9.58 19.96 8.77 23.9 12.7 76.22 89.67 25.3 April 25.93 14.37 29.26 15.43 27.3 15.4 28.19 30.80 77.3 May 32.03 18.74 36.58 21.45 34.9 21.4 47.70 26.70 44.5 June 28.23 23.23 29.77 24.23 40.5 25.0 17.58 100.83 14.3 July 33.55 23.52 34.16 24.61 32.6 24.2 185.51 149.90 271.3 August 32.42 24.00 34.94 24.50 32.6 24.2 270.05 150.38 204.3 September 31.50 21.37 32.17 21.47 32.0 19.7 - 164.50 37.3 October 29.74 15.48 30.74 15.77 28.3 14.3 23.34 - 53.7 November 25.50 9.86 25.50 10.83 24.7 9.5 18.43 2.47 1.9 December 19.03 6.74 20.85 7.40 16.5 6.0 14.03 - 42.1 Source: . Water Resources Research Institute, NARC, Islamabad 125 Table 9A. Agro-meteorological data at Gujranwala during, 1992, 1993 and 1994. Latitude 32° 10' N Longitude 73° 50' E Average Temperature (°C) Rainfall (mm) Month 1 992 1993 1994 1992 1993 1 994 Max Min Max Min Max Min January 20.1 8.3 19.0 6.7 20.6 8.3 61.1 10.6 24.0 February 21.1 9.7 24.6 13.2 21.7 9.8 32.1 14.9 18.8 March 26.7 14.4 25.6 13.3 29.3 16.4 12.0 40.1 5.6 April 32.7 19.5 34.5 20.1 32.3 18.7 16.5 31.8 9.8 May 37.4 23.1 40.7 26.1 39.9 25.5 24.9 8.0 36.7 June 40.9 26.8 40.2 27.9 41.7 28.5 17.5 28.3 13.6 July 35.5 26.8 35.0 26.5 35.8 28.1 88.0 182.9 128.6 August 34.5 26.8 37.7 28.7 34.0 27.0 196.4 33.5 154.1 September 34.4 24.6 34.2 24.8 34.0 23.8 150.8 24.3 115.5 October 33.8 18.9 31.3 18.2 32.3 17.9 5.8 - 4.3 November 27.6 13.2 28.9 13.1 28.1 13.5 9.7 0.5 0.5 December 23.7 9.5 23.8 7.6 21.7 8.9 7.5 - 27.5 Source: Water Resources Research Institute, NARC, Islamabad. 126 Table 10A. Agro-meteorological data at Multan during, 1992, 1993 and 1994. Latitude 30° 05' N Longitude 71° 40' E Average Tem perature (°C) Rainfall (mm) Month 1992 1993 1994 1992 1993 1994 Max Min Max Min Max Min January 20.6 6.2 20.4 5.9 21.2 5.9 19.0 11.4 5.2 February 22.4 8.6 26.5 9.7 22.2 7.2 16.1 1.5 6.4 March 27.3 12.9 27.2 12.5 31.3 15.0 6.0 12.7 5.2 April 25.7 33.1 36.0 20.0 34.2 18.1 16.7 34.5 3.2 May 39.8 23.6 42.6 28.7 41.7 25.5 8.6 26.1 39.8 June 36.1 28.5 42.1 28.6 43.6 29.4 1.5 3.2 Trc. July 38.9 28.5 36.9 27.7 37.5 28.2 15.5 209.5 51.3 August 36.1 27.4 37.6 28.0 36.7 28.2 217.3 1.8 18.6 September 34.1 24.0 37.1 28.3 34.4 23.6 201.5 Trc. 159.4 October 33.2 18.5 35.0 17.7 33.1 17.4 1.0 - - November 27.0 12.2 29.2 14.4 28.8 13.6 12.2 Trc. Trc. December 23.3 8.9 24.4 6.2 22.0 7.7 Trc. - 15.1 Source: Water Resources Research Institute, NARC, Islamabad. Trc. Trace 127 Table 11A. Estimated Potential Areas for the cultivation of Oilseed Crops 128 “Pakistan Edible Oilseeds Industry”, USAID. March 1984. Pakistan Economic Survey. 1998-99. in Pakistan. Land Description Area under Percent (%)area Fallow area different crops readily available available for (‘000’ha) for oilseed crops Oilseeds crops (‘000’ ha) Rice 2424 30 727.2 Cotton 2923 30 876.9 Rainfed 6000 25 1500 Riverine 4500 75 337.5 Inter-cropping 1155 25 288.75 (sugar cane) Total 3730.35 Source: Table 12A. Map of Pakistan Pakistan —— WWW --- mun-rm t 04mm 0 WW‘ w --—-- M wummw-‘IMAW m hymen-IQ 0::me 7“""1 '1 . in A4 I J i- F... ; BALOClilSI-AN/F~\ 129 Table 13A. Area, production and yield of soybean in Pakistan 1980 — 1999. YEAR Area Harvested Production (Mt) Yield (Kg/Ha) (Hal 1980 3512 1326 378 1981 3162 1342 424 1982 3691 1535 416 1983 4101 1350 329 1984 4465 1571 352 1985 4457 1602 359 1986 5446 2585 475 1987 5980 3775 631 1988 2758 1526 553 1989 2269 1169 515 1990 1495 849 568 1991 1875 930 496 1992 2193 1327 605 1993 4177 2355 564 1994 6613 5268 797 1995 6013 7228 1202 1996 2132 2694 1264 1997 5649 7311 1294 1998 6880 8500 1236 1999 8100 10000 1235 Source: . FAO Database 1980-99 (www.FA0.0r9.) 130