MSU RETURNING MATERIALS: P1ace in book drop to LIBRARJES remove this checkout from 4—,; your record. FINES wiH be charged if book is returned after the date stamped below. RELATIONSHIPS BETWEEN HORPHOLOGICALIAND PHYSIOLOGICAL CHARACTERISTICS AND YIELD OF DRY BEAN (PHASEOLUS VULGARIS L.) CULTIVARS DIFFERING IN THEIR PLANT ARCHITECTURE BY Joseph Michel Tohme 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 1986 ABSTRACT RELATIONSHIPS BETWEEN HORPHOLOGICAL AND PHYSIOLOGICAL CHARACTERISTICS AND YIELD OF DRY BEAN (PHASEOLUS VULGARIS L.) CULTIVARS DIFFERING IN THEIR PLANT ARCHITECTURE BY Joseph Michel Tohme The relationships between morphological and physiological characteristics and yield were investigated in seven sets of F6 lines near-isogenic for plant architectural traits and five parental cultivars of dry bean (Phaseolus vulgaris L.). The genotypes were classified as type I bush bean, type II architype beans, type II single sten beans and Type III beans. The twenty-two genotypes were tested for yield and morphological traits. A principal factor analysis and a modified principal component distance were performed on the morphological traits. Dry seed weight from the central three nodes on the main stem were collected on a four to seven days basis. Seed filling duration and rate were calculated for each genotype by fitting a cubic polynomial for the change in dry seed weight with time. A remobilization factor was calculated between mid-seed-filling and physiological maturity. Joseph Michel Tohme Four of the architype genotypes were the highest yielding genotypes and outyielded the conventional bush bean cultivar by almost 30%. The architype F6 lines outyielded their respective near-isogenic type I and type III lines. The factor analysis suggested that two different sets of associations exist for the determinate and the architype genotypes. It is inferred that the existing genic balance in the conventional cultivars has been replaced in the architype genotypes by a more efficient one. The genotypes had different remobilization factors for the period between mid-seed-filling and physiological maturityu The architype genotypes flowered later than the type I genotypes and had a longer seed filling duration when compared to their respective near-isogenic F6 lines. The seed filling rate was positvely correlated with seed size and negatively correlated with seeds per pod and days to maturity. The seed filling duration was correlated with yield and days to maturity suggesting that a short reproducyive period might result in a reduction in yield. ii To Lina and Riad ACNMNHEDGWMHS I wish to express my love and affection for my parents. The sacrifices they endured during my graduate studies allowed me to pursue the career I selected with complete freedom uninhibited by financial concerns. I am deeply grateful to Dr. Wayne Adams for his kind encouragement and his many helpful suggestions through the different phases of this study. His dedicated guidance and patience along with his in depth explanation of plant breeding concepts have made my study at MSU a truly beneficial experience. I am also grateful to Dr. Jim Kelly for encouraging my interest in plant breeding. His valuable comments throughout my study are appreciated. His challenging thoughts and his sharing of ideas have helped shape many aspects of my research. I am also grateful to the other members of my committee, Dr. Barbara Sears for enhancing my understanding of genetics, and Dr. Frank Dennis for making many helpful suggestions and for agreeing to serve on my committee on such a short notice. iii I would like to express my sincere appreciation to Dr. Russ Freed for his support and for providing me with both a working experience and employment. I am also grateful to Ms. Betsy Bricker, Dr. Scott Eisensmith, Ms. Mary Lauver and Mr. Jerry Taylor for their valuable assistance. I would like also to thank the many students who have given me their friendship and from whom I have learned a lot. Finally, I wish to acknowledge the major contributions of two dear friends, Lina Cortas for her active encouragement and moral support before and during the first months of my stay in the US, and Riad Baalbaki for his constant advice, his unselfish help, his faith and for being always available when needed. Without their support I could have never started nor completed my doctoral studies. iv TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES INTRODUCTION CHAPTER 1. CHAPTER 2. FACTOR ANALYSIS OF SEVEN FAMILIES OF NEAR-ISOGENIC F6 LINES AND FIVE CULTIVARS OF DRY BEAN DIFFERING IN THEIR PLANT ARCHITECTURE Abstract Introduction Materials and Methods Parental and Near-Isogenic Materials Planting and Harvesting Procedures Principal Factor Analysis Results Discussion References DRY MATTER REMOBILIZATION AND SEED FILLING PARAMETERS AMONG SEVENTEEN F LINES AND FIVE CULTIVARS OF DRY BEAN DIFFERING IN THEIR ARCHITECTURAL TRAITS. Abstract Introduction Materials and Methods Planting and Sampling Procedures Statistical Analysis Results Discussion References SUMMARY AND CONCLUSIONS APPENDICES Appendix A Appendix B Page vi viii 16 16 18 19 21 27 34 57 57 59 64 64 66 68 72 79 89 93 93 96 CHAPTER 1. 10. 11. 12. LIST OF TABLES List of parents used to generate the near- isogenic lines List of crosses made to generate the near- isogenic lines List of the F5 near-isogenic lines used in the study List of traits measured and estimated Mean yield of five dry bean cultivars and seven sets of near-isogenic F6 lines Means of various architectural traits of five dry bean cultivars and seven sets of near-isogenic F6 lines Factor loadings of the four most important factors for all the genotypes Factor loadings of the three most important factors for all the determinate genotypes Factor loadings of the five most important factors for all the architypes Distances between the parental genotypes based on seventeen morphological traits Distances between the determinate genotypes based on seventeen morphological traits Distances between the architype genotypes based on seventeen morphological traits vi Page 37 37 38 39 40 41 43 44 45 46 46 46 CHAPTER Appendix 1. 2 List of the twenty two genotypes of dry bean used and means of their primary yield components Days to 50% flowering, days to maturity, harvest index and yield efficiency of twenty-two genotypes of beans Stem dry weight at mid-seed-filling ($1), at physiological maturity ($2), seed dry weight at mid-seed-filling (G1) , at physiological 'maturity (62), and remobilization factor (RF) of‘twenty-two genotypes of dry beans between mid-sead- filling and physiological maturity (g/m ) Seed filling period (linear and effective) and seed filling rates of twenty-two genotypes of dry beans Correlation between length of seed filling, rates of seed fill and primary yield components and days to maturity for twenty-two genotypes of dry beans Regression parameter estimates and coefficients of determination of twenty- two genotypes of dry beans for the cubic polynomial regression equation for the seed dry weight upon times inldays after 50% flowering A Scores of twenty-two genotypes from the factor analysis of the morphological traits the factor analysis Scores of the determinate genotypes from the factor analysis of the morphological traits the factor analysis Scores of the architype genotypes from the factor analysis of the morphological traits the factor analysis vii Page 81 82 83 84 85 86 93 94 95 LIST OF FIGURES Page CHAPTER 1 1. Distribution of the genotypes on the basis 47 of their scores (”1 the first three principal factors, (a-c) plots for all the genotypes, (d-f) plots for the determinate genotypes, @rdn plots for the architype genotypes. CHAPTER 2 2. Seed filling curves of the cultivars 87 Swedish Brown, Seafarer and Swan Valley APPENDIX B 1. Seed filling curves of the twenty-two 96 genotypes. INTRODUCTION The manipulation of morphological traits has resulted in major yield gains in cereals (wheat and rice), and in legumes (bean and soybean). Plant breeders have improved yield by modifying plant architecture. The morphological traits manipulated are usually associated with important physiological processes. Such a strategy requires that breeders, using their knowledge of the physiology of a crop, and with the aid of statistical and mathematical modeling, define the ideal architectural plant for a specific environment. The concept of breeding for plant architecture involves the identificationtof morphological traits known to influence yield, and their combination into one genotype. In dry bean (Phaseolus vulgaris L.), an ideotype strategy has been proposed to improve yield of small seeded beans (Adams 1973L.The proposed plant model involved the restructuring of the bean plant using several architectural traits. Recent development of morphologically modified small seeded bean cultivars have incearsed yield potential by almost 30%. The new cultivars, similar to the model proposed by Adams, differ drastically from the conventional bush bean plant grown in Michigan under monoculture cropping system. These cultivars were classified as "architypes", a term coined by Adams (1982) to refelct their distinct architectural characteristics. As a group, the architypes are characterized by an indeterminate type II growth habit, a reduced number of branches, a narrow, erect profile, large number of seeds per pod, numerous pods and a long seed filling period. Existing variability and mutation of several architectural have been used to obtain the architypes. The genetic of some of these traits have been studied. However, little is known about the degree to which some of these traits may be associated and their relationships with yield. Physiological quantification for the differences in yield between the conventional and the architype cultivars is also lacking. Genetic differences has been observed for the remobilization of stem reserves. Under short period of stress, such remobilization would be of great importance for seed filling and yield. The study of the relationships between seed filling parameters and yield of the architypes has just begun. Some evidence has been accumulating recently suggesting that the length of the seed filling period is correlated with yield. The relationships between architectural traits, seed filling parameters and yield are investigated in this study. Six parental genotypes differing in their plant architecture were used to generate near-isogenic lines for morphological traits. Seven sets of near-isogenic F6 lines for plant architectural traits were obtained and tested for yield. The association of the morphological traits is examined by principal component analysis. Dry matters remobilization factors between mid-seed-filling and physiological maturity were calculated. Cubic polynomial equations is used to fit the data of seed dry weight change with time. The curves were then used to obtain seed filling parameters (duration and rate). The main objectives of this study was to investigate the relationships between morphological traits andjyield, and to identify associations among these traits and to offer a physiological quantification for yield by examining remobilization of stem reserves, seed filling period and rate of seven sets of near-isogenic lines for plant architecture and five cultivars of bean. CHAPTER 1 FACTOR ANALYSIS OF SEVEN FAMILIES OF NEAR-ISOGENIC F6 LINES AND FIVE CULTIVARS OF DRY BEANS DIFFERING IN THEIR PLANT ARCHITECTURE ABSTRACT Six accessions of dry bean (Phaseolus vulgaris LJ exhibiting differences in their growth habits were crossed to generate near-isogenic lines for plant architectural traits.'The parental genotypes included two determinate type I bush beans, two indeterminate type II single stem beans and two indeterminate type II architype beans. Seven sets of near-isogenic F5 lines were obtained from F4 families segregating for stem termination and architectural traits and were selfed to obtain seventeen F6 lines of small white seeded beans. Five parental genotypes and the seventeen F6 lines were tested for yield, primary yield components and fourteen various morphological traits. A principal factor and a principal component distance analysis were performed on the seventeen traits for the overall data, the determinate and the architype genotypes. Evidence is presented.to show that the seed yield of the F6 lines and the parental navy genotypes was associated with their modified architecture habits. All the high yielding genotypes were type II architypes. The F6 architype lines outyielded their respective near-isogenic type I«or type III lines. The principal factor and the principal component distance analyses suggested that two sets of different associations exist for the determinate and the architype genotypes. Several yield promoting traits and primary yield components were associated in the first factor for the architype but not for the determinate genotypes. The architype genotypes represented an optimum.genic balance more efficient under Michigan conditions than the one existing in the traditional determinate genotypes. INTRODUCEHMI Productivity of a crop results from the proper combination of genotype, environment and cultural practices. The complexity of the plant characters involved has been a major challenge and an uneasy task for plant breeders in search for higher yield. Yield depends, to a certain extent, on the functioning of numerous physiological and biochemical processes and their interaction with the environment. Many of such processes can be modified through the science of plant breeding and genetics. IDEOTYPE BREEDING According to Donald (1968), most plant breeding has been based on "defect elimination" and "selection for yield". He suggested that a third approach would be the breeding of crop ideotypes, plants with model characteristics known to influence photosynthesis, growth and grain production. Ideotypes could be seen as an intellectual and conceptual construction of a variety before any attempt was made at plant breeding. Such a breeding strategy required that the breeder define the environment for the model, identify morphological and physiological traits known to influence yield performance in that environment, formulate the ideal architectural plant, combine the traits into one plant type, and finally modify it upon testing (Adams, 1982). Since Donald's paper, there has been much interest and some skepticism among plant breeders concerning the potential of such a concept for yield improvement. Frey (1971) proposed the development of optimum plant types through the modification of yield components, morphological and/or, physiological traits. A limited number of plant breeders have endorsed and applied the concept, and from these several ideotypes have been defined for a limited number of crops; wheat (Donald, 1968), barley (Donald, 1979), rice (Jenning, 1964'), corn (Mock and Pearce, 1975), peas (Davies, 1981), and dry beans (Adams, 1973). Conventional plant breeders have been reluctant to emphasize individual characters or several characters that might constitute an ideotype (Rasmusson and Gengenbach, 1983). Since primary yield components are correlated with yield, many attempts have been made to utilize primary yield components as selection criteria in improving grain yields often, however, with little success (Nickel and Grafius, 1969; Coyne, 1968). This failure to obtain satisfactory progress from selection based on yield components has been attributed to yield component compensation, an almost universal phenomenon expressed in the negative correlations between component characters which develop in a sequential pattern (Adams, 1967). The primary yield components in beans, (number of pods, number of seeds per pod and seed size), are to some extent interdependent in their development. Since the interdependce is generally negative, the effects of any increase in one yield component are off- set by decreases in one or more of the other components. BREEDING FOR PHYSIOLOGICAL TRAITS The value of morphological, physiological or biochemical traits depends on several criteria, among them the existence of genetic variability for trait expression, the knowledge of its genetic control, and its relationship to agronomic benefit (Mahon, 1983). Less progress has been made in selecting for specific physiological and biochemical traits associated with.yield.]kngeneral, the results of research directed towards improved photosynthesis have been disappointing as avenues to production of improved varieties. Sufficient information has been lacking as to which physiological or biochemical characters influence yield and the optimum phenotype for these characters. Wallace et a1. (1972) reviewed the work of many scientists and concluded that genotypes within species do indeed exhibit variation for most physiological components of yield. Furthermore, they proposed the breeding of beans for higher photosynthetic rate. But later studies failed to show any association between high assimilation rate and yield (Kueneman et al. 1979). Direct selection for photosynthesis or related characters has not established the existence of any consistent effect on growth or dry matter yield (Wilson, 1981). The growing understanding of crop physiology as attested by several studies (Wallace et al., 1972, Evans and Wardlaw, 1976, Wilson, 1981, Evans, 1983) should eventually permit the design of crop plants with potentially greater yield than existing cultivars. However, until the physiological and biochemical basis of yield are better understood, identification and development of physiological and biochemical selection tools useful to plant breeders will be delayed. BREEDING FOR HORPHOIDGICAL TRAITS Another current strategy shared by some breeders interested in yield improvement lies in the exploitation and the modification of the plant architectural traits associated with physiological function. Frey (1971) expressed more optimism about improving yield capacity of crop plants through selection for certain morphological types as compared to selection of physiological traits. Assisted by a better understanding of the factors involved, plant breeders have made significant advances in yield 10 breeding by their manipulating of plant habit of soybeans (Cooper, 1974), beans (Kelly ettal”.1984), wheat and rice (Davies,1983). It is apparent that morphological traits have been modified considerably through selection. Most of these traits are readily observable and quantifiable, and can therefore easily be selected in plant breeding programs. As plants were domesticated, selection for or against certain features took.place, and plant architecture was modified. The more common problem is determining the best combination of traits, which may be inter-correlated, and the feasibility of combining them into single cultivars for a specific cropping system. Reddy and Sinha (1971) cited several architectural features as having a pronounced effect on crop growth. Among these traits were: plant type as influenced by the canopy and thickness of leaf, leaf area, and light penetration as affected by the display of leaves. Other architectural features include the number and arrangement of branches or tillers (Donald, 1968). When dealing with morphological traits, the variability and genetics of the trait are of prime importance to the plant breeder; Ghaderi and Adams (1981) found quite high broad sense heritability for such traits as plant height, nodes above 15 cm., hypocotyl diameter, number of pods, number of seeds per pod and seeds per plant in dry beans. 11 BREEDING FOR HORPHOIDGICAL TRAITS IN GRAIN LEGUHES In grain legumes, Adams (1982) has formulated the basic principles and fundamental architectural objectives for achieving maximum yield potential. The plant architecture should maximize light interception, while the canopy profile should consist of as many source-sink units as possible. The adjustment of the source-sink ratio should maximize the sink size relative to the source. In soybeans, Cooper (1974) demonstrated that lodging may be the most important single limitation to exploitation of yield potential in higher yielding environments. Through the modification of soybean plant architecture by genes for stem growth habit and maturity, several semi-determinate high yielding varieties have been developed (Cooper, 1985). In peas, the so-called semi-leafless peas.retain.improved standing ability without causing a serious growth reduction due to insufficient photosynthetic area. As a result, several semi-leafless varieties are being grown commercially in the United Kingdom (1977). In both cases, as in the case of the semi-dwarf'wheat, the plant habit.has been changed dramatically by simple manipulation of traits controlled by a few genes. However, such an impact on yield improvement is usually associated with an important effect on physiological processes. 12 BREEDING FOR HORPHOIDGICAL TRAITS IN DRY BEANS In beans (Phaseolus vulgaris I»), considerable genetic variation exists for plant architectural traits. Singh (1982) classified the growth habit of beans into four major plant types (type I, II, III, and IV), with sub-division for types II, III and IV. The type I exhibited a determinate growth habit, while type II, III and IV exhibit an indeterminate growth habit. These three indeterminate types differed in the length of the vine growth, branch angle, and the canopy structure with types II being more upright than types III and IV. In general, the growth habit is affected by the environment, temperature and photoperiod. In addition to the natural variability in the bean germplasm, several architectural traits such as determinate type I, type II and single stem types have been obtained through mutation plant breeding (Adams 1982). Factor analysis has been used in biological sciences to study the relationship between several traits (Walton, 1971, Denis and Adams, 1978). Principal factor analysis is a mathematical technique for reducing a large number of correlated variables into a smaller number of patterns of variables called factors. The derived factors are independent of each other with the first factor accounting for the largest proportion of the variation. It allows one 13 to observe some underlying patterns of relationships that might exist among several variables by reducing the number of factors accounting for the variance. Each factor is a hypothetical variable that contributes to the variance of at least two of the observed variables. Denis and Adams (1978), using factor analysis on 22 traits of 16 pure line bean cultivars at two locations, identified patterns of morphological characteristics. Factor I was identified with the number of reproductive structures while factor II was identified with the size of the reproductive structures. Based on several morphological and physiological studies, Adams (1973, 1983) has specified a new architectural plant type drastically different from the conventional navy bean plant grown under a monocultural cropping system. The new plant would be a tall, narrow profile indeterminate bean plant with a dominant main stem and two to four basal but erect branches. The stem must be a dominant central axis and must have as many nodes as possible. The upper internodes must be longer and more numerous than the basal internodes. Leaf size should also be small and capable of orientation to allow optimal light penetration through the canopy. The pods should be long with many present at each node and with six to seven seeds per pod. 14 Evans (1973) questioned the value of the bean ideotype on the grounds that it is "dogmatic and unrealistic" to breed for too strict a morphological ideotype for any particular defined environment. Instead, she favored giving great priority to, and breeding for, plant adaptability rather than designing a model suited only to a particular environment. Coyne (1980) evaluated plant architectural modification and stated that insufficient information on the contribution and the relative merits of many of the morphological and physiological yield components is available to develop a model which is likely’tijroduce a high yielding plant. Instead, he advocated conventional breeding methods such as selecting parents with superior level of morphological and physiological traits associated with yield and using the parents in breeding programs with other high yielding germplasm. In an attempt to study the effect of plant architecture on beans, Singh and Gutierrez (1979) developed several lines with enhanced or suppressed morphological traits. Their lines had one of the following architectural traits: plant with less than 10 nodes on the main stem, erect branching, suppressed branching, short foliage and internodes, lanceolate leaves, small pods and long pods. Some of these lines were compared with four conventional cultivars at 16 plant density-environment combinations (Nienhuis and Singh, 1985). None of the newly constructed 15 architectural lines yielded more than their respective check varieties in any environment or at any plant density. Despite the initial skepticism (Evans, 1973), several navy bean lines, conforming to the initial principles stated in Adams's paper (1973), have been constructed and several varieties have been released in Michigan since 1981. These varieties, called "architypes", have been tested and proven to be superior in yield to the existing conventional varieties over locations and seasons (Kelly et al., 1984). The main objectives of this study were to develop navy bean lines near-isogenic for plant architectural traits, with similar genetic background, using various parental lines differing in their plant architecture. These lines and the parental genotypes were compared for archictectural traits and yield, to identify important patterns of traits and to determine whether there is a different association among the patterns identified for the determinate and architype genotypes. MATERIALS ANDIHETHODS Six dry bean accessions were used to generate sets of near—isogenic lines with the same genetic background but differing in plant architectural traits. The parents, listed in table 1, were chosen to represent a wide range of variability in terms of plant architectural characteristics such as growth habit, number of branches, number of nodes, leaf area, number of seeds per pod and seed size. Parental and.Near-Isogenic Materials The accessions described according to CIAT classification under Michigan environmental conditions and cultural practice consisted of: 1) Seafarer, a determinate navy type I bush bean which has been the earlier maturing standard cultivar, 2) Swedish Brown, a large yellow seeded type I, 3) Swan Valley, a narrow profile, high yielding type II navy bean, 4) MSU experimental line 790458, a narrow profile type II navy bean, 5) MSU experimental line 61319 a single stem type II navy bean and 6) MSU experimental line 791515, a single stem type II white seeded bean. Seafarer and Swedish Brown were selected for their 16 17 different origin of determinancy. Swan Valley and 790458 were chosen as representatives of the architype lines, characterized by a narrow profile with a central dominant stem: plants are tall, erect, non-lodging high yielding navy bean cultivars.IMSU experimental lines 61319 and 791515 were chosen for their main and dominant single stem. The crosses made are listed in table 2. The F1 seeds were planted in the field in East Lansing, Michigan during the summer of 1982. Around 250 F2 seeds from each cross were sent to Isabela, Puerto Rico, where selections were made in February, 1983. Only F2 plants showing intermediate growth habit were selected, thus establishing around 25 F3 families from each of the crosses. The F3 plants were planted in the field at East Lansing during the summer of 1983 and additional selections were made between and within families for indeterminate heterozygous lines. Around 10 heterozygous lines from different F3 families were sent to Isabela, Puerto Rico, during the winter of 1983-1984. Selections were made within the F4 families for plants differing in their architecture. The characters used in the selection procedure were determinate versus indeterminate and/or architype versus viny or single stem plants . The F5 seeds were increased at Fortuna, Puerto Rico, to produce F6 lines for the study. Although not all desired combinations for the architectural traits were found, several different near-isogenic lines 18 were obtained and are listed in table 3. Lines selected from the same F4 family will be referred to as lines ~having the same genetic background. Planting and Harvesting Procedures Five parents and the seventeen F6 lines were planted in June 12, 1984 at East Lansing, Michigan, using a precision drill mounted air planter. Due to the lack of seed, the parental accession 61319 was not planted, from past record its performance should be similar to line 791515. Each experimental unit consisted of 8 rows S‘m long, spaced 50 cm. apart. Within-row spacing was 7-8 cm giving 13-15 plants per meter of row. The experimental units were arranged in a randomized block design with four replications of twenty-two plots each. Standard practices of herbicide and fertilizer application were used. Irrigation was applied two weeks after planting to compensate for the deficiency in natural rainfall during that period. Uniformly spaced plants were harvested at 50% flowering and at the normal physiological maturity for each line. Data were recorded for individual plants and all the architectural characters measured and calculated are listed in Table 4. Yield data were obtained by harvesting the middle two rows and seed weight was reported at 14% seed moisture content. 19 Principal Factor Analysis A principal factor analysis from the Statistical Analysis System (SAS) package was performed on the means of the variables listed in table 4 for the overall data set, the determinate sub-set and the architype genotypes. Factor analysis established a set of derived factors which are independent of each other and fewer in number than the original variables. If two correlated variables with a number of measured scores are plotted, the contour of the scatter diagram will form an ellipse. The two axes or vectors defining the ellipse are the principal factors. The long axis is the first factor, accounting for as much of the total variance as possible and the other axis is the second factor accounting for the remainder of the total variance. Mathematically, the concept can be extended for any number of variables in a multi-dimensional space. The projection of the variable vector onto the factor is called the factor loading coefficient and constitutes the correlation between that variable and the factor. The steps involved in the analysis included the use of the correlation matrix for all 17 variables. The prior communality for each variable was set to its maximum absolute correlation with any other variable. The factor loadings were extracted from the eigenvalues and eigenvector matrices. To make the interpretation less subjective the 20 factors were rotated using an orthogonal Varimax rotation, that is, by applying a nonsingular linear transformation. The transformation was used in order to establish orthogonality among the factors. The interpretation was accomplished by examining the loading coefficients of variables in each factor, with particular attention to coefficients in the range of 0.5 to (L9. A conceptual name was then assigned to each factor that reflects the importance and the biological meaning of the factor. In order to illustrate the relationships among the various genotypes, the scores of the genotypes for the first three factors were plotted in two dimensional graphs with factor I, II and III as axes. A principal component distance program for the SAS package was used to calculate distances between the genotypes based upon the morphological traits listed in table 4. The program was designed to calculate distances among a set of bean cultivars by Adams and Wiersma (1977). Rather than using the D2 statistic of Mahalanobis, the distances are calculated from the normalized principal component scores using the equation: dij = [0:11 - le) + H. (xik - xjk)2]1/2 where di’ equals the distance between varieties i and j on J principal axes I through k, and X11 is the normalized score of variety 1 on axis 1. RESULTS The various F6 lines, with the exception of line 16, expressed the architectural traits for which they were selected (Table 4). Although line 16 was selected for the single stem trait, at the planting density used, it showed a reduced number of branches instead. The F6 architype lines were very close to the architectural structure of the released architypes, but perhaps not as refined. All the F6 lines were white-seeded and, based on their seed size, could be considered to belong to the navy bean class. The means for seed yield and for various architectural traits are presented in tables 5 and 6. Significant differences among the genotypes (cultivars and the F6 near- isogenic lines) were observed for seed yield and the architectural traits both within and between genetic backgrounds. Genotypes with type I growth habit were earlier maturing, shorter, with more branches and fewer nodes on both the main stem and the branches than either the type II or type III progenies. Seafarer and Swedish Brown, on the average, matured several days earlier than the F6 lines. The F6 lines classified as architypes matured two to four days earlier than the architype parents. Seafarer and Swedish Brown were the shortest stature genotypes and had the most 21 22 numerous branches, whereas line 790458 was the tallest genotype and had the lowest number of branches. YIELD OF THE PARENTAL GENOTYPES AND THE F6 LINES. The yield of the genotypes ranged from 1961 to 3318 kg/ha., where Swan Valley was the highest yielding genotype and the single stem 791515 as the lowest yielding genotype. The already established architypes (Swan Valley and 790458) and the F6 architype lines 2 and 13 significantly outyielded the conventional navy bean cultivar Seafarer. The increase in yield over Seafarer was 29%, 23%, 29% and 27% for Swan Valley, 790458, and F6 lines 2 and 13, respectively. The four architype lines also outyielded the single stem cultivars by around 40%. The increase in yield for these genotypes over Seafarer and 791515 was significantly different at the 1% level (Table 5). Swedish Brown, a large seeded bean (40.1 g/100 seed) was higher yielding than Seafarer. Its yield, though numerically less, 'was not significantly different from the top four yielding architypes (LSD .01 = 482 kg/ha). In the genetic background of families.1, 2 and 6, the F6 type II architypes outyielded their respective near- isogenic type I or type III lines by 23%, 19% and 29%, respectively. The increase in yield was statistically significant at the 1% level. In the genetic backgrounds of 23 families 3, 4, and 5, the architype F6 lines were higher yielding than their near-isogenic lines by at least 10%, although this difference was not statistically significant. In the genetic background of family 7, consisting of type I, a bush type II, and a reduced branching type II, no significant differences in yield were observed. The type III F6 lines were similar in yield both to their near-isogenic type I lines and to the type I lines in different genetic backgrounds. Between the different genetic backgrounds, F6 lines 2 and 7 were higher in yield than any other F6 line. No type I F6 line produced a higher yield than the traditional navy bean Seafarer. The F6 line 5, a determinate architype, had a lower yield than its near- isogenic type II architype, although the difference was not statistically significant. GENETIC DISTANCE AMONG THE GENOTYPES AND FACTOR ANALXSIS OF ALL GENOTYPES, THE TYPE I AND THE ARCHITYPE GENOTYPES. The factor loadings for the complete data set (cultivars and F6 lines), and for the determinate and the architype genotypes, are listed in tables 7, 8 and 9, respectively. The loading coefficient represents the correlation between a trait and a particular factor. Only the traits with the underscored high values of the loading coefficients were used in defining the factors. Major 24 emphasis will be put on the factor interpretation of the determinate and architype analysis, whereas the factor for the overall data set will be used to extract the factor scores of each line. For the complete data set (twenty-two genotypes), four main factors, accounting for 78% of the total variance, can be recognized. The four factors account for 31.5%, 23.0%, 13.8% and 10.0% for the factors I, II, III and IV, respectively. The first factor included the number of branches, branch length at 50% flowering and maturity, number of nodes on the branches, branch internode length and average long internode length on the main stem.lNegative loading consisted of plant height, number of nodes on the main stem and number of nodes above 10 cm., all being main stem traits. The second factor included plant height, number of nodes on the main stem, number of nodes below 10 cm, number of pods and total number of nodes. The negative loading coefficients were for seed size and short internode length, indicating negative compensation between the primary yield components. The third factor included only seed size: its negative loading coefficient for number of seeds per pod and average long internode length also indicate a compensatory relationship. The number of factors and the loading coefficients for the various traits were quite different for the determinate 25 and the architype genotypes (Tables 8 and 9). For the determinate type, three factors emerged (Table 8), accounting for 83% of the total variation, with two factors almost equivalent in respect to the variance for which they accounted. The first factor included plant height, total branch length, total number of nodes and nodes on branches. The negative loading coefficient for seed size indicated yield component compensation. The second factor included number of branches and branch internode length. A strong negative loading coefficient for number of seeds per pod also indicated a strong component compensation among primary yield components. The third factor included number of pods and number of nodes on main stem. The factors in the architype genotypes (Table 9) were dissimilar to the factors for the determinate genotypes.'The first five factors, accounting for 91% of the total variance, will be examined. The first factor was dominated by number of pods, seeds per pod, number of nodes on branches and number of nodes on main stem. Negative loading included branch length, branch internode length and average long internode length on the main stem. The second factor included number of nodes below 10 cm., and number of nodes on branches.'The third factor included mainly seed size with a weak negative loading coefficient for number of pods. The fourth factor included number of nodes on branches, and plant height at 50% flowering. The fifth factor was 26 dominated by the number of branches. The graph of the first two factors for the overall analysis (Figure 1) indicated that the parental genotypes were quite divergent from each other, with the notable exception of Swan Valley and 790458. This observation is supported by the calculated genetic distances among the parental genotypes (Table 10). Seafarer and 791515, the single stem genotype, were the farthest apart (Distance = 1.3063), whereas Swan Valley and 790458 were the closest (Distance = 0.3663) . Further examination of the various graphs (Figures 1- 6), and the calculated distances among the genotypes (Table 11) indicated that the determinate genotypes were also divergent from the type II parents, but even more from the determinate parents. Line 5, the architype determinate genotype, was the most divergent from the other type IIF6 lines. While all the F6 architypes were quite divergent from either one of the determinate parental genotypes, most of them were quite close to’ either one of the architype parental genotypes (Figure 7-9 and Table 12). Surprisingly, line 2 was closer to 790458 than to its parental architype Swan Valley. The determinate line 5 was quite divergent from its F6 architype near—isoline and its parental genotypes. DISCUSSION The morphological data collected were used in the factor analysis for the determinate and the architype genotypes and will be discussed in this context. It should be stressed that the results and interpretation that emerge are completely dependent upon the data used and no physiological interpretation will be made.'The choice of the data set was a deliberate one, in order to identify the morphological differences between the conventional type I and the type II architype beans. The first factor in the type I genotypes (Table 8) is a composite factor that includes both vegetative vigor, in the traits of plant height and branch length, but also potential reproductive structures in the form of number of nodes and nodes on branches. There is a certain logic in their being loaded in the same factor, since the number of nodes and number of nodes on branches should be greater when the branch number and length are greater. The second factor completes the vegetative aspect of the first factor with the number of branches and branch internode length. The factor also possesses a negative dimension due to the compensatory relationship between plant height, branch and hypocotyl diameter. The third factor is mainly a reproductive factor 27 28 with the number of pods being heavily loaded. Also heavily loaded is node number on the main stem and number of nodes above 10 cm. The first factor for the architype genotypes (Table 9) is a major reproductive structure factor, an architectural factor in which number of reproductive structures is of the predominant importance. The negative loading for branch length is a consequence of the expression of negative vegetative component compensation. Such compensation is in agreement with findings of Harmsen (1984) who showed that any increase in the main stem will result in a decrease in the branches. The second factor completes the yield promoting factor, and can be identified as a vegetative vigor and reproductive potential factor. The factor promotes long branches with numerous nodes on branches. The third factor is a seed size or weight factor. The branch internode length is not important to the identity of the factor. The loading is only fortuitous for the traits and it should not be thought to characterize or identify the factor. The negative loading for number of pods is a consequence of the expression of a weak yield component compensation. The remaining two factors are basically vegetative vigor factors with the number of branches being loaded heavily on the fifth factor. The last factor suggests that the number of branches is not as important as the branch length. 29 Based on the overall factor analysis (Table 7 and Fig. 1) and the calculated genetic distances among the genotypes (Table 10-12), the genotypes used in this study could be defined in terms of several gene pools, where each gene pool represents a sample of genes interacting to produce a genic balance and thus an acceptable array of genotypes for the Michigan environment. The determinate navy genotypes consisted of two gene pools represented by the F6 lines and Seafarer, respectively, whereas the indeterminate architype consisted of one gene pool. The determinate F6 lines were quite different from the traditional cultivar Seafarer and from the architype parents which confirms visual observation that the attempt to recover the parental type I was not fully successful.‘The difference among the determinate lines and Seafarer is not surprising. It is the direct result of the method of enforced heterozygosity used to generate the near isogenic F6 lines. The indeterminate F6 architypes were more similar to either one of the architype parents, than to each other. Further examination of the main loading coefficients of the various factors and of the calculated distances suggests that two different sets or associations of genes exist for the determinate and the architype genotypes. Each association of genes, relatively independent of the other, regulates a pattern of related.growth.or‘development of‘a number of traits. In the architype genotype several yield 30 promoting traits and primary yield components are associated in the first factor, whereas in the determinate genotypes the association exists only for potential reproductive traits. Such associations have broad implication in breeding for plant morphological architecture. The existing genic balance among the various traits in the traditional determinate genotypes has been replaced by a more efficient or potentially higher yielding one. While the various architectural traits individually can have an influence on yield, it is their integration into a particular genic balance that will eventually result in an increase in yield. The parental architypes are a representative of such an optimum genic balance. The bean breeders at Michigan State University have been able to transfer several important morphological traits and combine them into one genotype. The first factor in the architype analysis (Table 9) indicates that the various traits promoting yield have been associated together in contrast with the determinate genotypes. It also appears from the number of determinate architype lines obtained during the selection procedure that it is more difficult to obtain a determinate architype line by a simple cross followed by selection. Dry bean yield increases in Michigan have been relatively low for several years when compared to 31 increases in other field crops. Advances have been made in disease resistance, and in cultural practices. It was not until recently that increased grain yield pg; gg has been achieved mainly by restructuring the bean plant and by combining several physiological and morphological characteristics into one genotype. The use of near-isogenic lines offered an opportunity to further test the relationship between yield and plant morphological modification in the navy bean plant, grown under Michigan conditions. The near-isogenic lines developed did not include all possible combinations in one genetic background nor did they attain the refined morphological canopy of some of the released lines. However, the lines obtained allowed comparison among several different architectural plants within the same genetic background and among several genetic backgrounds. The seed yields of the various F6 lines and the architype parents were associated with their modified architectural growth.habit (Table 5). Line 2 and 13 outyielded significantly the conventional variety Seafarer and their respective near-isogenic lines. All the superior lines were type II architypes, with a longer growing season and seed filling period. They outyielded significantly the conventional navy bean Seafarer and their near isogenic type I and III lines. These results do not agree with the finding reported by 32 Nienhuis and Singh (1985), who did not find any morphologically modified genotypes of any growth habit capable of outyielding the commercial cultivars. Furthermore, indeterminate prostrate type III and type II cultivars in that order were among the highest yielding, regardless of environment and plant densityu It should be noted that the various lines used in their study each had a pronounced expression of one of the various architectural traits rather than a combination of several traits. The developmental interdependency among the various architectural traits is well documented.(Adams 1982). It would be difficult in the bean plant to achieve an increase in yield by simply manipulating one or few architectural traits. The high yielding F6 lines obtained in this study were closely similar to the conceptual model offered by Adams (1982). The architype cultivars released in navy and black beans have been consistently superior in yield to the conventional small-seeded type I genotypes across locations and years. The architypes are also able to capitalize on a longer seed filling period associated with a larger sink (Izquierdo and Hosfield 1983). The breeding of architypes has required simultaneous selection for several traits at once and.a willingness to compromise in the expression of a specific trait. The new 33 architypes arose by bminging together genetic variability from the Central American small seeded black beans with domestic navy types, thus permitting a remodeling of architectural growth habit through selection. Such a remodeling happened by using an ideotype strategy. The bean ideotype defined by Adams was, however, never intended to be a universal one. The bean architypes were designed specifically for the Michigan environment and a high density planting arrangement. Rarely are two cultivars identical. They may share the same gene pool and follow a similar strategy to maximize yield, but the development of the sequential traits of seed yield is usually different. The present study pointed to the advantage of the morphologically modified genotype over the traditional bush navy bean type. As a group, the F6 architype lines obtained share the same gene pool as their parental type II architypes. They are characterized by an indeterminate growth habit with a longer growing season, a narrow tall profile, with reduced branching and numerous nodes and seeds per pod. As a group, their morphological traits important to yield are associated together in contrast to the case of the type I bush navy beans. It should be noted that the type II architypes represent in no way a model to be used under an environment markedly different from the Michigan agro-ecological environment, without extensive testing. REFERENCES Adams, M.W., 1967. Basis of yield component compensation in crops plants with special reference to the field bean, Phaseolus vulgaris L. Crop Science 7:505-507. Adams, M.W., 1973. Plant architecture and physiological efficiency in the field bean. In "Potentials of Field Bean and Others Food Legumes In Latin America". Series Seminars No. 2B. Centro International de Agricultura Tropical, Cali, Colombia. Adams, M.W., 1981. The new bean "Uprights", their characteristics, performances and management. Bean Commission Journal, Nov page 20-22. Adams, M.W., 1982. Plant architecture and yield breeding. Iowa State Journal of Research 56(3): 225-254. Adams, M.W. and J. Wiersma, 1977. An adaptation of principal components analysis to an assessment of genetic distance, with an example. Research Report, Technical Information. Mich. Agric. Expt. Stn. E.Lansing, Michigan. page 1-7. Cooper, R.L., 1972. Influence of early lodging on yield of soybeans. Crop Science 11: 449-450. Cooper, R.L., 1978. Elf: A new semi-dwarf soybean variety for high yield environments. Illinois Research 20(4): 3-40 Cooper, R.L., 1981. Development of short statured soybean cultivars. Crop Science 21: 127-131. Coyne, D.P., 1968. Correlation heritability and selection of yield components in field beans, Phaseolus vulgaris L. Proc. Am. Soc. Hortic. Sci. 93:393-396. Coyne, D.P., 1980. Modification of plant architecture and crop yield by breeding. Hortscience 15: 244-247. Davies, R.D., 1977. Restructuring the pea plant. Sci. Prog. Oxf. 64:201-214. Denis, J.C. and M.W. Adams, 1978. A factor analysis of plant variables related to yield in dry beans. I- Morphological traits. Crop Science 18: 74-78. Donald, C.M., 1968. The breeding of crop ideotypes. Euphytica 17:385-405. 34 35 Donald, C.M., 1979. A barley breeding program based on an ideotype. J. Agric. Sci.,Camb. 93: 261-269. Evans, A.M., 1973. Commentary upon plant architecture and physiological efficiency in the field beans. In "Potentials of the Field Beans and Other Food Legumes in Latin America" page 279-286. Series Seminars No. 2E. Centro International de Agricultura Tropical, Cali, Colombia. Evans, IuT. and I.F. Wardlaw, 1976. Aspects of the comparative physiology of grain yield in cereals, Advances in Agronomy, 28:301-359. Evans, L.T., 1983. Raising the yield potentialzby selection or design. In:"Genetic Engineering of Plants" page 371-389, Kosuge, T., Meredith, C.P. and A. Hollander, (eds), Plenum Press, New York. Frey, KJL, 1971. Improving crop yield through plant breeding. In R. Munson, (ed), Moving off the yield plateau. Spec. Publ. No. 20, Amer. Soc. Agron, Madison, Wi, page 15-58. Ghaderi, A. and M.W. Adams, 1981. Preliminary studies on the inheritance of structural components of plant architecture in dry beans. Bean Impr. Coop. (New York) 35-37. Harmsen, ILH., 1983. Genetics and developmental inter- relationships of architectural traits in beans (Phaseolus vulgaris) with emphasis on branching. M.Sc. Dissertation. Michigan State University, East Lansing, Michigan. Kelly, J.D., M.W. Adams A.W. Saettler, G.L. Hosfield and A. Ghaderi., 1984. Registration of C-20 navy bean. Crop Science 24:822. Kueneman, E. A., D. H. Wallace, and PAM. Ludford, 1977. Photosynthetic measurements of field-grown dry beans and their relation to selection for yield. J.Amer. Soc. Hort. Sci. 104:480-482. Jennings, ILR., 1964. Plant type as a rice breeding objective. Crop science 4:13-15 Izquierdo, J.A. and G.L. Hosfield, 1983. The relationship of seed filling to yield among dry beans with differing architectural forms. J.Amer. Soc. Hort. Sci. 108:106- 111. 36 Mahon, J.D., 1983. Limitations to the use of physiological variability in plant breeding. Can. J. Plant Sci. 63: 11-21. Mock, J.J. and R.B. Pearce, 1975. An ideotype of maize. Euphytica, 24: 613-623. Nickell, C.D. and J.E. Grafius, 1969. Analysis of a negative response to selection for high yield in winter barley, H; vulgare. Crop Science 9: 447-451. Nienhuis, J. and S.P. Singh, 1985. Effects of location and plant density on yield and architectural traits in dry beans. Crop Science 25:579-584. Rasmusson, D.C. and 8.6. Gengenbach, 1983. Breeding for physiological traits. In "Crop Breeding". American Society of Agronomy, Madison, Wisconsin, page 231-254. Reddy, B.V.S. and S.K. Sinha, 1971. Genetic control of photosynthetic activity and its importance in plant breeding, a review. Indian J. of Genetics and Plant Breeding. 31:94-104. Singh, S.P., 1982. A key for identification of different growth habits of Phaseolus vulgaris L. Bean Impr. Coop. (New York) 25:6-8. Singh, S.P. and J.A. Gutierrez, 1982. Source of some architectural traits in dry bush beans. Bean Impr. Coop. (New York) 25:8-10. Wallace, D.H., J.L. Ozbun, and H.M. Munger, 1972. Physiological genetics of crop yield. Advances in Agronomy 24: 97-147. Wardlaw, I.F. 1968. The control and pattern of movement of carbohydrates in plants. Bot. Rev. 34: 79-105. Wilson, D., 1981. Breeding for morphological and physiological traits. In K.J. Frey (ed),"Plant Breeding II". Iowa State Univ. Press, Ames, Ia, page 233-290. 37 Table 1. List of parents used to generate the near-isogenic lines. Lines MSU Accession # Plant type Architecture Seafarer N67001 I Standard Bush Swedish Brown 181099 I Bush Swan Valley N76004 II Architype 790458 N79021 II Architype 61319 N76010 II Single Stem 791515 X80004 II Single Stem Table 2. List of crosses made to generate the near-isogenic lines. Cross # Cross 1 Seafarer x Swan Valley 2 Seafarer x 790458 3 Seafarer x 61319 4 Swan Valley x 791515 5 Swan Valley x Swedish Brown 6 790458 x 791515 7 791515 x Swedish Brown 38 Table 3. List of the F6 near-isogenic lines used in the study. F6 Line F4 Cross Plant Plant # Family Type Characteristics 1 1 Seafarer x Swan Valley I Determinate 2 1 Seafarer x Swan Valley II Indet. Architype 3 2 Seafarer x 790458 II Architype 4 2 Seafarer x 790458 III Indet. Viny 5 3 Seafarer x 790458 I Det. Architype 6 3 Seafarer x 790458 II Indet. Architype 7 4 Seafarer x 61319 I Det. Bush 8 4 Seafarer x 61319 II Indet. Architype 9 4 Seafarer x 61319 III Indet. Viny 10 5 Swan Valley x Swedish Brown I Det. Bush 11 5 Swan Valley x Swedish Brown II Indet. Architype 12 5 Swan Valley x Swedish Brown III Indet. Viny 13 6 790458 x 791515 II Indet. Architype 14 6 790458 at 791515 III Iniet. Viny 15 7 791515 x Swedish Brown I Determinate Bush - 16 7 791515 x Swedish Brown II Indet. Red. Br. 17 7 791515 x Swedish Brown II Indet. Bush 39 Table 4. List of traits measured and estimated. (DIE'IRAITS 10. 11. 13. 14. 15. 16. 17. PL'IHI‘ Plant height PIIIHI‘SO Plant height at 50% flowering NUMNODE Total number of nodes NODDB Number of nodes on the main stem NODEHR Ntmber of nodes on branches NIMBR Number of branches mo Branches lergth at 50% flowering BRIGIH Branches length HYPDIAM Hypoootyl diameter NUMPOD Number of pods SEEDPOD Nunber of seeds per pod SEEIBIZE Seed size NODEAIO Ntmber of nodes above 10 an. mosaic Number of nodes below 10 an. AINI'LBR Average internode length on branches AIIND‘ISAveragelonginternodemthemainstem ASIN'DB Average short internodeonthemainstem Table 5. Mean yield of five dry bean cultivars and seven 40 sets of near-isogenic F6 lines. F4 F6 Lines number Growth Plant Yield @ Family and Genotypes Habit Architecture (Kg/ha) 1 1 Seafarer x Swan Valley I Determinate 2523 bode 1 2 Seafarer x Swan Valley II Indet. Architype 3295 a 2 3 Seafarer x 790458 II Architype 2860 abc 2 4 Seafarer x 790458 III Indet. Viny 2304 de 3 5 Seafarer x 790458 I Det. Architype 2208 e 3 6 Seafarer x 790458 II Indet. Architype 2508 bode 4 7 Seafarer x 61319 I mt. Dish 2257 de 4 8 Seafarer x 61319 II Indet. Architype 2781 abcd 4 9 Seafarer x 61319 III Iniet. Viny 2229 de 5 10 Swan Valley x Swedish Brown I Det. Bash 2479 bode 5 11 Swan Valley x Swedish Brown II Indet. Architype 2993 ab 5 12 Swan Valley x Swedish Brown III Inlet. Viny 2528 bode 6 13 790458 x 791515 II Indet. Ardiitype 3205 a 6 14 790458 x 791515 III Iniet. Viny 2273 de 7 15 791515 x Swedish Brown I Determinate Bush 2113 e 7 16 791515 x Swedish Brown II Indet. Single Stan 1968 e 7 17 791515 x Swedish Brown II Irdet. Bash 2275 de Olltivar SEAFARER I Det. 31511 2326 Ode Cultivar SWEDISH m I Bet. Bush 2883 ab Cultivar SWANVALIEY II Architype 3318 a Exp. Line 790458 II Architype 3030 ab Exp. Line 791515 II Single Stem 1961 e ISD (0.01) 481.8 CV (’6) 10.0 0 Mean followed bythe same letter or letters are not significantly different according to mncan's Multiple Range Test at 1% level. 41 mm c._~ m.m o.- m.¢~ w.q o.n~ m3.~m no Om ELL :3oum gnapwam x >m~am> scam m mm m.- 0.5 o.o~ n.q~ _.m m.m_ mo.qm no as dd :30um cmapozm x zufifim> :mBm m Nm ~.NN «.0 ~.m~ o.n~ o.o o.~_ mw.oc an ac _ shown :mwtmzm x >m~Hm> cmzm m mm ~.NN «.0 R.LN ¢.m~ o.¢ m.m~ Om.oo «a me -~ m~m~c x hummumom q on m.o~ o.e N.mm ~.n~ m.q ~.N~ mm.mm «a me ~_ o~m_o x umummmum e mm m.- m.o ~.@~ w.m~ w.< ~.m mm.qv cm 06 _ odmdo x hummumom q on o.w~ m.o «.0N o.m~ m.m m.m~ mo.wn we me _~ mncoou x umumwmwm m Em ~.n~ ~.m o.mm m.- 5.6 N.@ Mm._m mm me E wmqomb x uwummmmm m mm ¢.q~ 5.6 ¢.qm m.m~ o.m o.o~ oo.om «a be -~ mncomm x uuummmum N mm m.m~ m.o m.~m N.~m m.q c.q~ mc.cm mm cc - mm¢9¢n x umumwoom m on m.o~ N.n N.oN n.o~ o.c m.m~ 0@.No cm me dd >w-m> swam x umpmmmwm A mm ~.m~ m.o N.ON m.- m.q o.¢ mo.oc om Ne E >w-m> :m3m x umumwmom a A.uwv A.Eov A.SUV madam \ uuwm com \ ucmam \ cameo: muzucmum Boom can: osmuo: zowusomz weauozofim been: xafismm mmvoz Lo boom meowm meom we Locmum wo “@2832 so mocoz Sagan cu mxma Cu mzmo nuaouo moamuocmo em umnasz ponfizz wmmpo>< owmum>< .mmcua on o«cuwom« use: Co muum cm>mm ecu mum>fiudzo ammo xuo o>aw mo mufimuu Hmu3uoouficoum wacfiug> mo was»: .0 mange 42 ow o.m~ ~.e m.- o.o~ ~.N m.w~ mm.m~ mm Ac HE wamfimn mafia .exm mm m.o~ m.m ~.n~ m.o~ ~.c m.m~ mn.~o mo on Efi wmqoon wow; .axm om c.m_ ~.n o.om n.m~ m.q n.o~ ma.mm no me - >m44<> z<3m um>wufizo om ~.oe «.6 o.- m.n~ m.o c.w mo.mm mm mm E 230mm zmuanm um>fiuazo dc o.n~ ¢.m ~.o~ w.m~ m.m o.o mo.me Em mm L mmdauasu mm ~.m~ m.o N.m~ o.w~ ~.m m.o~ m~.mm on me - caoum smumoaw x man—mm m mm ~.o~ m.m ~.o~ m.¢~ ~.q m.- mm.mm om om EH Sacha smupoam x mdmfimn m «N m.o~ m.m ~.w~ m.om m.q 0.x Om.mq mm mm a caoum smapozm x mdmdmn m on w.o~ m.o ~.o~ e.m~ o.q m.m~ o~.mn mm on Eda n~n~a~ x mmeoam e um m.o~ o.m ~.w~ «.mm m.m m.m~ mm.mm oo~ Om dd m~m~o~ x mmeoom 0 A63 753 753 ucmfim \ ouwm pom \ ucmfim \ nuwsoq monocsum Scum can: ucwwmz xuausumz wcuuo3on Dana: >Husmm mmeoz Lo poem mpoom mean we possum mo nonaaz so mopoz ucwfim cu m>ma cu mama nuzopo mmazuocoo em bonsaz bonasz ammum>< ammuo>< .mocfia on casewoma may: we mama cu>om com mum>qu~=o ammo sup u>uw mo mufimuu _ms:uovowzoum macaum> mo mama: .Aoucoov 0 dance 43 for all the W. Table 7. Factor loadings of the four most important factors TRAITS FACIORl FACIORZ memes FACIOR4 PLEHI' -0.66035 0. 50764 0.27624 -0.28480 PIIIHI‘SO -0.16889 0. 31087 -0.43927 -0.29536 NUMNOI'E 0. 33751 0.74472 0. 21859 0. 30542 NODEMS -0.64240 0. 62883 0.40322 -0.13643 NODEHI 0 . 69294 0 . 40585 -0 . 00088 0 . 28287 NUME? 0.78618 -0.12583 0.41315 0.34694 HUGH-£50 0.77601 0.09522 -0.13107 -0.13586 BRIGIH 0.75235 0.22482 0.34085 -0.30344 HYPDIAM -0.62 669 0. 18238 -0.44166 0.06423 NUMPOD 0.26787 0.57828 0.13547 0.46773 SEEDIOD -0 . 00559 0 . 54872 -0 . 45496 0 . 46424 SEEIBIZE -0.29179 -0.61490 0.55201 0.06255 NODEA10 -0 . 74107 0 . 49648 0 . 41241 -0 . 07828 MIG 0.41961 0.66235 -0.07324 -0.43642 AINTIR 0. 59778 -0. 19437 0. 58329 -0. 35878 ASIN'DB -0. 43153 -0. 68608 0. 11927 0. 52444 AL'INDB 0 . 53930 -0. 35572 -0. 55783 -0. 32349 PROPORTION 0.3157 (IJMJIATIVE 0.3157 0. 2301 0.5458 0. 1382 0. 6840 0. 1006 0. 7846 44 Table 8. Factor loadings of the three most important factors for all the determinate genotypes. TRAITS FACIORl m FACTORS PIEHI‘ 0. 67768 -0 . 71430 0. 08936 PUIHI‘SO 0 . 21975 -0 . 61542 -0 . 49883 NUMIDDE 0 . 75999 0 . 59382 0 . 08310 NODE‘B 0 . 53183 -0 . 12928 0 . 82682 NODEH? 0. 71125 0 . 65356 -0 . 06025 NW 0. 27895 0 . 94381 0 . 06110 HUGH-150 0 . 82443 -0 . 19709 -0. 01994 mam 0. 68518 0. 53670 -0. 39381 HYPDIAM -0. 48171 -0. 73883 0. 15307 NUMPOD 0 . 28588 0 . 12505 0 . 80959 SEEDPOD 0 . 12759 -0 . 55184 0 . 10527 SEEIBIZE -0. 76223 0 . 46518 0 . 26372 NODEAlO 0. 09811 -0 . 03691 0. 87796 NODEB10 0 . 87111 -0 . 38057 0 . 11477 W 0. 35155 0 . 78364 -0 . 32234 ASIN'DB -0 . 90398 0 . 39889 -0 . 09838 ALINDB 0 . 45674 -0 . 53210 -0 . 59901 PROMO)! 0.3488 0.3052 0. 1839 GJMJIATIVE 0.3488 0.6540 0.8379 45 Table 9. Factor loadings of the five most important factors for all the architypes. TRAITS m FACIORZ FACIOR3 FACIIR4 FACTOR 5 PL'IHT 0. 78010 0. 21720 -0. 15021 -0. 27210 -0. 03310 PIER-H50 0. 01129 0. 38919 -0.07668 0. 53643 -0.70639 NUMNODE 0 . 63090 0 . 63569 -0 . 21765 0 . 34488 -0 . 00372 NODEMS 0 . 80552 0 . 40012 0 . 19001 -0 . 36681 -0 . 07211 NOEBR 0.24405 0. 52270 -0. 39890 0. 67630 0.04387 NUMHR 0 . 02083 0 . 34557 -0 . 00055 0 . 44010 0 . 73550 MGR-150 -0 . 83650 0 . 19651 -0 . 16238 -0 . 39160 0 . 10565 W -0 . 47950 0 . 59707 0 . 62540 0 . 10344 0 . 01265 HYPDIAM 0 . 33027 -0 . 27997 0 . 00168 0 . 46314 0 . 40095 NUMPOD 0.64145 0.31853 -0.34634 -0.39426 0.41119 SEEDPOD 0 . 89440 -0 . 14829 0 . 24011 0 . 16988 -0. 13525 SEEIBIZE -0. 06560 0. 08703 0 . 94745 0. 22803 -0. 00693 NODEAlO 0.88143 0.22896 0.20092 -0.33716 -0.07720 NO-10 -0 . 50757 0 . 76055 -0 . 23366 -0 . 21789 -0 . 04868 AINTIH? -0. 48718 0.41355 0.72711 -0. 08904 0.20374 ASINDB 0 . 46123 -0 . 77975 0 . 27009 0 . 17255 0 . 07500 AIINDB -0. 84630 -0 . 22515 -0 . 43332 0. 16720 0. 00594 HQOPORI'ICN 0.3630 0.1894 0. 1558 0.1240 0.0861 (IMJIATIVE 0.3630 0.5524 0.7082 0.8322 0.9183 46 Table 10. Distamesbetweentheparentalsgenotypesbasedm seventeen morphological traits. S . Brown S . Valley 790458 791515 Seafarer 1 . 0727 1 . 1062 0 . 9502 1. 3063 S. Brown 1.1112 1.0737 1.1592 8. Valley 0.3663 1.1040 790458 0 . 9706 Table 11. Distances between the determinate genotypes based on seventeen morphological traits. Line 5 Line 7 Line 10 Line 15 Seaf. S.B. S.V. 790458 791515 Line 1 0.6054 0.4174 0.5089 0.4788 0.9780 0.9794 0.7890 0.7034 0.9932 Line 5 0.7910 0.7930 0.9177 1.0472 1.1766 0.9813 0.8465 1.1321 LIne 7 0.4434 0.5018 0.9448 0.8538 0.9980 0.6684 1.0079 Line 10 0.6922 0.8341 0.6950 0.5467 0.5401 0.9945 Line 15 0.9060 1.0761 0.9487 0.8384 0.9044 Table 12. Distances between the architype genotypes based on seventeen morphological traits. Line 3 Line 5 Line 6 Line 8 Line 11 Line 13 S.V. 790458 Line 2 0.9326 0.7204 0.2956 0.4513 0.6020 0.3921 0.5013 0.2225 Line 3 1.1600 0.7508 0.8668 0.8167 0.8041 1.0398 0.8733 line 5 0.5924 0.8618 1.0501 0.9830 0.9813 0.8465 Line 6 0.5190 0.6395 0.4591 0.6134 0.3417 Line 8 0.6372 0.4175 0.4304 0.4293 Line 11 0.5544 0.8248 0.5955 Line 13 0.5592 0.3469 47 Figure 1. Distribution of the genotypes on the basis of their scores on the first three principal factors, (a-c) plots for all the genotypes, (d-f) plots for the determinate genotypes, (g-h) plots for the architypegemtypes. (SF=Seafarer, SB=Swedish Brown, SV=Swan Valley, A2=790458, SS=791515) . 48 no :DTJ—K3)>*1 o mm mm 36.3.» a .a—J N-w .. «(yo_ 0 so = + as. s Figure la. 49 DJ JUC>-KD3>W1 ~L ‘1 mm 320.53 E d— N—w “CE 5. even. 0 .53 = + .76.. :_ Figure lb. 5.0 mm ma .3 LA ZDCJ—K33-T1 C) 3296“” M b .966. o .360 = 1. din :— Figure 1c. 51 mm mm NJ JUCJ—K3)P*1 O c 1>0403 E m... I mafia-ow mm I 93%.... was: r .76.. Figure 1d. 52 u woqo>w o me I. c 1.8.0.3.» ._ ‘- » mmurggfi MGIQEFSQSS > £2. Figure 1e. 53 (A IDC>—K3)”fi C3 .me mm l e 2.08» n n mm..m99€:= mm.um¥aa?:_wdta 5. demo. Figure 1f. 54 >~ N moao>m T ow< c 390403 E w m< I men: (2.3. >M l 363.5 0.. l e 0 >335!— u m 55 Cd JUC>-K3)"1 d1 I. c “50402 ._ d-J m< I mic: <95. 5 I «@950 0 2.07360 Figure 1h. 56 OI 20CP4CN>”W CI c .9040» M m< I men: <95. >~ I .3050 0 26:35.0 Figure 1i. CHAPTER 2 DRY MATTER REMOBILIZATION AND SEED FILLING PARAMETERS AMONG SEVENTEEN F LINES AND FIVE CULTIVARS OF DRY B ANS DIFFERING IN THEIR ARCHITECTURAL TRAITS. ABSTRACT The purpose of this experiment was to examine the inter-relationships between dry matter remobilization, seed filling parameters and plant architecture in seventeen.F6 lines and five cultivars of dry bean. The genotypes used represented a wide range of variability with respect to days to flowering, days to maturity and plant architectural characteristics. Thirty days after flowering, dry weight of stems, leaves, pods and seeds were measured for each genotype. From these data, the change in stem, grain dry weight and a remobilization factor were determined between mid-seed-filling and physiological maturity. After 50% flowering, dry seed weights from the central three nodes on the main stem were collected on a four to seven days basis. The data were fitted for each genotype using a cubic polynomial model.‘The linear seed filling period (LFP), effective filling period (EFP) and linear filling rate (LFR) were calculated. 57 58 The genotypes differed in their accumulation and remobilization of stem reserves. No consistent trend was found between the different genetic backgrounds, suggesting a different sink demand or a lesser ability of some genotype to re-allocate reserves. All the high yielding lines had a long seed filling period. On the average the type II architypes had a relatively longer linear filling period than type I or type III genotypes. The two seed filling parameters were highly correlated with each other which would indicate that the linear filling period alone could be used to evaluate the length of the seed filling period. The linear filling rate was negatively correlated with.days to maturity; The seed filling period parameters were correlated with yield and days to maturity suggesting that a reduction in the reproductive period might result in a reduction in yield. INTRODUCTION The adoption of an ideotype strategy for improvement of a grain legume requires careful selection of useful morphological traits. These morphological traits must be associated with physiological processes related to sink development. High yield is only achieved through the proper combination of numerous physiological and morphological components in each genotype. However, each cultivar attains yield through its own combination of these components and its interaction with environmental factors. Thus, stable yield promoting traits must be identified and combined into one genotype. The efforts in achieving high yield in beans (Phaseolus vulgaris) under Michigan conditions have been concentrated on development of architype cultivars. These architypes are characterized by an indeterminate growth habit, a narrow profile, small leaves capable of orientation under strong light, a large number of potential reproductive sites, a high remobilization of dry matter, a long seed filling period and a high rate of seed filling (Adams, 1983). Partitioning Among Plant Parts Plant breeders have devised various indirect selection 59 60 criteria in developing strategies for increasing yield. Direct selection for photosynthesis or related.characters has not established the existence of any consistent relationship with growth or yield of dry matter (Wilson, 1981). In wheat, modern high yielding varieties produce no more biomass than older varieties of comparable growing period (Austin et al., 1980). In bean, Tanaka and Fujita (1979) found no significant cultivar effect upon photosynthetic rate in leaves of comparable age. Harvest index has been widely used to express the relationship of partitioning of dry matter between the biological and economical yield of plants. Due to the difficulties of obtaining an accurate measurement of root dry weight, only shoot dry weight is usually considered. Thus the harvest index has been defined as the ratio of economic product, such as grain, to the above ground biomass at harvest. Donald and Hamblin (1976) have reviewed the use of the concept of harvest index. They pointed out that when harvest index is used as the sole parameter for selection, it may fail to predict the yield of a crop. It does not take into consideration differences in total biomass. In both barley and wheat, increase in grain weight often exceeds the dry weight of the vegetative portion of the plant, sometimes by 50% or more (Gallagher et a1” 1975). In these instances, some portion of grain growth is 61 apparently sustained by translocation of assimilates produced before anthesis and stored temporarily in the stem. Grain legumes in general and beans in particular have been described as a series of phytomeric or source-sink units (Adams and Pipoly, 1980). Each unit consists of a raceme associated with a leaf, node and internode. Normally, a major fraction of the assimilate in the seed unit will have been produced in the leaf associated with the same phytomeric unit (Dure, 1975). Such association is, however, never complete and movement from adjacent phytomeric units may occur, depending on the ratio of source to sink. Some evidence suggests that.yields in