lllllllllllllllllHINHHIUIIIIHllllll\IHUUIIIHUIHHI LIBRARY 301404 308 MiChigan Sm: University This is to certify that the thesis entitled THE INHERITANCE OF SEVERAL CHARACTERISTICS IN CAULIFLONER AND PEPPER presented by DENNIS J. WERNER has been accepted towards fulfillment of the requirements for Ph.D. Horticulture degree in flIT/ucc A1574” Z: Major professor Date March 30, 1979 0-7639 OVERDUE FINES ARE 25¢ PER DAY PER ITEM Return to Book drop to remove this checkout: from your record. THE INHERITANCE OF SEVERAL CHARACTERISTICS IN CAULIFLOWER AND PEPPER By Dennis J. Werner A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Horticulture 1979 ABSTRACT THE INHERITANCE OF SEVERAL CHARACTERISTICS. IN CAULIFLOWER AND PEPPER BY Dennis J. Werner Studies were carried out on cauliflower to develop a measurement and sampling technique for the determination of leaf geometry, and to investigate the inheritance of this character. A measurement technique was developed which suggested that a valid measure of a plant's leaf geometry was obtained by using the mean of the first, fifth, and ninth leaves subtending the curd. Data from two field plantings showed that upright leaf geometry was partially dominant, with both additive and dominance gene action involved in trait expression. Evidence was obtained to suggest that epistasis was involved in the expression of leaf geometry. Narrow-sense heritability estimates of 67 percent and 65 percent were obtained for the two field plantings. Expected gain from selection estimates of 31.5 percent and 27.8 percent were also calculated. Genetic analysis suggested that 3 major loci (A, g, 9) controlled the expression of leaf geometry in cauliflower. A (9:7)(3zl) factorial of the 3 loci best explained the observed ratios. The presence of both A and §_dominant genes were essential for upright leaf expression, i.e., Dennis J. Werner §_and §_were complementary to each other. 9 was epistatic to the ex- pression of a3 and by, the presence of dominant §_in combination with either A- or B- conditioning upright expression. Thus, upright expres- sion can occur by the presence of the §_gene in combination with genes A or l_3_ or by the presence of A and E_ in the absence of _C__. Studies on peppers were conducted to determine the mode of inheri- tance of pedicel width and fruit-calyx detachment force, and to determine the relationship of these characters with fruit length, width, and weight. Data from 6 x 6 diallel analysis and F analysis showed that pedicel 2 width was quantitatively inherited, with additive and dominance gene action both involved in the expression of this character. Epistatic gene action was also suggested. Estimates of the number of effective factors, 5, indicated that approximately 12 loci were involved in control- ling pedicel width. Narrow-sense heritability estimates of 84 percent and 67.5 percent were obtained in the diallel and F experiments, respect- 2 ively. An expected gain from selection of 21 percent was estimated. Correlation coefficients between pedicel width with fruit length, width, and weight were calculated from both the diallel and F2 analysis. All correlations were significant at the 1% level in both experiments, but were of greater magnitude in the diallel analysis. Fruit-calyx detachment force was shown to be controlled by additive gene action, with estimates of §_ranging from 1.29 to 1.86. Correlation coefficients between detachment force and fruit length, width, and weight were calculated from F2 data. All coefficients were positive and significant at the 1% level. To my grandparents, Sylvia and Milton Bixler, for all they have given me, and To the memory of Mrs. Elsie Benedick, who taught me to appreciate the beauty of a flower. ii ACKNOWLEDGMENTS I wish to express my sincere appreciation to Dr. S. Honma for giving me the opportunity to pursue my graduate education under his guidance. I also wish to acknowledge the members of my guidance committee, Dr. Charles Cress, Dr. James Hanover, Dr. L. W. Mericle, and Dr. Robert Herner for their friendship and advice. To the numerous individuals who assisted in taking the data, my sincerest thanks and appreciation. Special thanks to Amos Lockwood, Fred Richey, "Louie" Pollok, and Bill Priest for their invaluable assistance in field-plot preparation and maintenance. To my parents and my wife, Georgina, for their love and support. iii TABLE OF CONTENTS Page LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . Vi LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . .Viii CHAPTER I: THE INHERITANCE OF LEAF GEOMETRY IN CAULIFLOWER (BRASSICA OLERACEA L. VAR. BOTRYTIS) INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . 1 REVIEW OF LITERATURE . . . . . . . . . . . . . . . . . . . . . 4 MATERIALS AND METHODS . . .’. . . . . . . . . . . . . . . . . 8 Development of Sampling Technique . . . . . . . . . . . . 8 Field Studies . . . . . . . . . . . . . . . . . . . . . . 17 RESULTS AND DISCUSSION . . . . . . . . . . . . . . . . . . . . 21 Development of Sampling Technique . . . . ... . . . . . . 21 Field Studies - Pooled Analysis . . . . . . . . . . . . . 23 Field Study (Summer) . . . . . . . . . . . . . . . . . . 23 Field Study (Fall) . . . . . . . . . . . . . . . . . . . 37 SUMMARY AND CONCLUSIONS . . . . . . . . . . . . . . . . . . . 46 CHAPTER II: THE INHERITANCE OF PEDICEL WIDTH AND FRUIT DETACHMENT FORCE IN PEPPER (CAPSICUM ANNUUM L.) INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . 52 REVIEW OF LITERATURE . . . . . . . . . . . . . . . . . . . . . 54 MATERIALS AND METHODS . . . . . . . . . . . . . . . . . . . . 56 Pedicel Width - Diallel Analysis . . . . . . . . . . . . S6 Pedicel Width — F2 Analysis . . . . . . . . . . . . . . . 58 Detachment Force . . . . . . . . . . . . . . . . . . . . 61 iv Page RESULTS AND DISCUSSION . Pedicel Width - Diallel Analysis . . . . . . . . . . . . 67 Pedicel Width - F2 Analysis . . . . . . . . . . . . . . . 73 Detachment Force . . . . . . . . . . . . . . . . . . . . 80 SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . 88 LITERATURE CITED . . . . . . . . . . . . . . . . . . . . . . . . . 90 Table 10. 11. 12. 13. 14. LIST OF TABLES Geometric values (cm) for individual leaf and various leaf combinations for MSU 839 . . . . . . . . . . . . . . Geometric values (cm) for individual leaf and various leaf combinations for MSU 831 . . . . . . . . . . . . . . . Analysis of variance for leaf geometry fOr summer and fall crop I I I I I I I I I I I I I I I I I I I I I I I I I I I Bartlett's test for homogeneity of variance on F2, BC;, and BC2 generations individually compared between the two planting times . . . . . . . . . . . . . . . . . . . . . . Frequency distributions of leaf geometry values (cm) in the cross MSU 831 x MSU 839 . . . . . . . . . . . . . . . Significance of the A, B, C and joint scaling tests for leaf geometry in a MSU 831 x MSU 839 cross . . . . . . . . Gene effects estimated using three and six-parameter models on generation means in a MSU 831 x MSU 839 cross . . Chi-square test for 3-gene model for leaf geometry in the F1, BC;, and BC2 in the cross MSU 831 x MSU 839 . . . . . . Proposed genotypes and phenotypes of F2 and backcross generation individuals . . . . . . . . . . . . . . . . . . Frequency distributions of leaf geometry values (cm) in the cross MSU 831 x MSU 839 . . . . . . . . . . . . . . . . Significance of the A, B, C and joint scaling tests for leaf geometry in a MSU 831 x MSU 839 cross . . . . . . . . Gene effects estimated using three and six-parameter models on generation means in a MSU 831 x MSU 839 cross . . . . . Chi-square test for 3-gene model for leaf geometry in the F1, BC;, and BC2 in the cross MSU 831 x MSU 839 . . . . . Analysis of variance of leaf geometry . . . . . . . . . . . vi Page 22 24 25 26 28 31 33 35 36 38 42 43 44 47 Table Page 15. Analysis of variance of leaf geometry . . . . . . . . . . . . 48 16. Characteristics of the parental lines. Pedicel width investigation . . . . . . . . . . . . . . . . . . . . . . . . 59 17. Characteristics of the parental lines. Detachment fOrce investigation . . . . . . . . . . . . . . . . . . . . . . . . 62 18. Mean pedicel width (mm) for a set of diallel crosses between 6 inbred lines of pepper . . . . . . . . . . . . . . . . . . 68 19. Analysis of variance of diallel table for pedicel width . . . 69 20. Components of variation for pedicel width . . . . . . . . . . 71 21. Correlation coefficients of pedicel width with fruit width, length, and weight . . . . . . . . . . . . . . . . . . . . . 72 22. Frequency distributions of pedicel width values (mm) in the cross MSU 333 x MSU 365 . . . . . . . . . . . . . . . . . 74 23. Significance of the A, B, C and joint scaling tests for pedicel width in a MSU 333 x MSU 36S cross . . . . . . . . . 77 24. Gene effects estimated using three and six-parameter models on generation means in a MSU 333 x MSU 365 cross . . . . . . 79 25. Correlation coefficients (r) and coefficients of determi- nation (r2) of pedicel width with fruit width, length, and weight in a F2 generation of the cross MSU 333 x MSU 36S . . 81 26. Frequency distributions of detachment force values (kg) in the cross MSU 160 x MSU 249 . . . . . . . . . . . . . . . 82 27. Correlation coefficients (r) and coefficients of determi— nation (r2) of detachment force with fruit width, length, and weight in the F2 generation of the cross MSU 160 x MSU 249 . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 vii LIST OF FIGURES Figure Page 1. Leaf geometry of the two parental types . . . . . . . . . . . 10 2. Longitudinal view of cauliflower plant, showing arrangement of leaves . . . . . . . . . . . . . . . . . . . . . . . . . . 12 3. Schematic representation of the various leaf types en- countered in cauliflower plants, showing absence of clearly defined angle between leaf base and main axis . . . . l4 4. Diagram of technique used in measuring leaf geometry in cauliflower . . . . . . . . . . . . . . . . . . . . . . . . . l6 5. Frequency distributions for leaf geometry (cm) in cauli- flower from the cross MSU 831 x MSU 839 . . . . .2. . . . . . 30 6. Frequency distributions for leaf geometry (cm) in cauli- flower from the cross MSU 831 x MSU 839 . . . . . . . . . . . 40 7. Instrumentation used to measure fruit detachment fOrce . . . 65 8. Frequency distributions for pedicel width (mm) in pepper from the cross MSU 333 x MSU 365 . . . . . . . . . . . . . . 76 9. Frequency distributions for detachment force (kg)in pepper from the cross MSU 160 x MSU 249 . . . . . . . . . . . . . . 84 viii CHAPTER I THE INHERITANCE OF LEAF GEOMETRY IN CAULIFLOWER (BRASSICA OLERACEA L. VAR. BOTRYTIS) INTRODUCTION Geometric configuration of leaves on a plant has important impli- cations relative to the cultural and production practices utilized. The alteration of leaf geometry also offers a possible means of modifying photosynthetic efficiency, and therefore, total yield. Several workers (Pearce et al., 1967; Sakamoto and Shaw, 1967; Tanner et al., 1966) have suggested that leaf angle is related to crop efficiency. Variation exists among cauliflower (Brassica Oleracea L. var. botrytis) germplasm in the geometric configuration of the leaves, thus presenting the plant breeder with the opportunity to utilize this variation in the develop- ment of new cultivars. Most cauliflower produced in Michigan is hand—tied, a practice necessary to present curd discoloration caused by sunlight exposure. Discolored or yellow curds are commercially unacceptable. Manual tying of the leaves is estimated to be 70 percent of total labor costs. Migrant labor is also becoming increasingly difficult to obtain during the fall when most of the crop is harvested, and many Michigan growers are reluctant to grow cauliflower. In an attempt to reverse this trend, efforts were initiated in the 1960's to develop cauliflower cultivars that did not require tying. These efforts were realized with the development of the cultivars 'Greenball' in 1971 and 'Self-Blanche' in 1973. 'Greenball' is char- treuse-green at maturity and does not require tying, while the 'Self- l 2 Blanche' cultivar produces inner leaves that wrap around the curd, pro- tecting the curd from sunlight. Although these cultivars are used commercially, there are certain disadvantages associated with them. Public acceptance of green cauliflower has been poor. For 'Self—Blanche', the wrapping character is manifested only when night temperatures are less than 10°C, therefore, during a warm growing season, manual tying of the leaves would be necessary. Problems with 'Self—Blanche' also arise if hot weather arrives after the leaves have wrapped around the curd. This causes a rise in the temperature around the curd due to in- sufficient aeration, and often results in riciness. Under conditions which favor heavy guttation, such as foggy conditions, guttated water released from the wrapper leaves onto the curd can cause brown spots, resulting from the salts present in the guttation water. Although 'Self-Blanche' and 'Greenball' made a significant contri- bution to no-tie cauliflower production, the disadvantages associated with these cultivars made it necessary to search for germplasm with alternative means of natural curd protection. This effort has resulted in the development of strains with other types of leaf arrangement for curd protection. Currently, strains are being developed in which the leaves assume a more upright or erect position around the developing curd, in contrast to a more horizontal position. The upright leaves form a 'stovepipe' arrangement around the curd, therefore, protecting it from sunlight. Besides providing natural sun protection to the curd, the upright leaf habit allows for sufficient aeration, preventing the problems associated with high temperatures. Information regarding the genetic control of leaf geometry would be helpful in designing an efficient breeding scheme for the development 3 of superior genotypes. The objectives of the following study were: 1. To develop a technique for measuring leaf geometry. 2. To develop a sampling technique appropriate for a genetic study. 3. To investigate the inheritance of leaf geometry. REVIEW OF LITERATURE Variation in leaf geometry exists in many agriculturally important plant species. The majority of the literature focuses on the effect of leaf geometry on plant growth and yield. Donald (1968b) suggested that erect leaves are more efficient in light utilization within a plant community than horizontal leaves. He bases this on the fact that erect leaves allow better light penetration into the crop canopy, allowing for adequate illumination of a greater leaf area, as compared with horizontal of drooping leaves, in which the upper leaves would be overlit and the lower leaves harmfully shaded. However, Anderson (1966) reported that upright leaves do not necessarily result in a uniform light distribution into a crop canopy, since light distribution is also a function of angle of the sun, cloud cover, crop density, wind, and other factors. A number of studies have demonstrated the value of erect leaves on crop growth rate and yield. Chang and Tagumpay (1970), in studying the association between grain yield and 6 agronomic traits in rice, found that high grain yield was significantly correlated with erect leaves. Tanner et a1. (1966) characterized approxi- mately 300 strains of barley, wheat, and oats, and fOund that erect leaves were associated with high yield in all cases but two. Similar conclusions have been reached by other workers (Gardener et al., 1964; Pearce et al., 1967; Pendleton et al., 1968; Rutger and Crowder, 1967). 5 In contrast, various studies have failed to detect any significant beneficial effect of leaf angle on yield. Yap and Harvey (1972) fOund that erect leaves in corn were not associated with high yield. However, these authors used a low planting density, and cautioned that the value of erect leaves is probably related to crop canopy parameters and grow- ing conditions. Studies by Hicks and Stucker (1972), Russell (1972), Ariyanayagam et a1. (1974), and FrOlich et a1. (1977) also failed to con- sistently support the beneficial effect of upright leaves on yield. The interrelationship of leaf angle and crop density to yield was investigated by Winter and Ohlrogge (1973), who found that at a high leaf area index (high crop density), upright leaves tended to increase grain yield in corn, whereas at a low leaf area index (low crop density), upright leaves reduced yield. Whigham and Woolley (1974), and Hopper (1971), using corn, also determined that upright-leaved types were superior in yield only when grown in narrow rows and/or at high popula- tion densities. Duncan's (1971) computer simulation model, which pre- dicts that at a high leaf area index, more erect leaves should give greater production. In contrast, model computations done by Kuroiwa (1970) showed that with canopies at optimal leaf area indices, the most efficient utilization of light will occur when the uppermost fOliage is erect, the lowest foliage horizontal, and the transitional foliage angles at intermediate levels. Although studies evaluating the role of leaf angle on crop produc- tivity have resulted in conflicting results, Mock and Pearce (1975), and Donald (1968a) still maintain that erect leaves are an important component in the development of ideotypes of corn and wheat, respectively. 6 Gardener et a1. (1966) discussed the relationship of crop leaf angle to weed growth. They Observed that at locations where weeds were chemically controlled, upright-leaved wheat strains were highest yield- ing. However, at locations with no herbicidal treatments, upright types yielded lowest. They observed that weed growth between the rows of the upright-leaved types was more profuse as compared with the horizontal- leaved types. The horizontal leaves apparently established a canopy over the rows and suppressed weed growth. These observations tend to support the findings of Sinclair (1972), and Allen (1973), who showed that in an erect-leaved wide-row maize canopy, more irradiation reached the soil surface than in a normal-leaved canopy. ‘ Literature on measurement of leaf geometry is limited. In crop plants, investigators have used direct measurement between the base of the leaf and the main axis of the plant using either a protractor (Ariyanayagam et al., 1974; Yap and Harvey, 1972) or a clinometer (Whigham and Woolley, 1974; Mason and Zuber, 1976). Genetic studies on leaf geometry are also limited. Seetharaman and Srivastava (1971) reported that erect flag leaf in rice was controlled by a single recessive gene. Yap and Harvey (1972) investigated the in— heritance of flag leaf angle in barley using a 7 x 7 diallel over 2 years, and found highly significant additive gene effects. In corn, combining ability estimates for leaf angle expression were equally influences by additive and non-additive gene effects (Mason and Zuber, 1976). Nigam and Srivastava (1976) investigated the inheritance of leaf angle in wheat and concluded that additive gene action controlled trait expression. SwarupznulChatterjee (1972) described the leaf geometry characteris- tics of the major European cauliflower types. These same authors (Swarup 7 and Chatterjee, 1974) classified the leaf geometric type of 290 inbred lines derived from inbreeding 200 heterozygous Indian cauliflower culti- vars. However, no genetic investigation was undertaken. MATERIALS AND METHODS Development of SamplingfiTechnique This investigation was undertaken to obtain a technique for measur- ing the leaf geometry of the cauliflower plant. Two inbred (6 genera- tions) cauliflower lines, MSU 831, a horizontal-leaved line, and MSU 839, an upright, were utilized (Figure 1). In cauliflower, the mature plant consists of a single axis, terminating in a curd which is subtended by leaves arranged in a whorl at various stages of development (Figure 2). Due to the whorled leaf arrangement and the absence of a clearly defined angle between the base of the leaf and the main stem (Figure 3), measure- ment with a protractor or clinometer was not feasible. Alternatively, a relative measure of geometry fOr an individual leaf was determined in the following manner (Figure 4): A point 35 cm from the base of the leaf petiole (B), measured along the midrib of the leaf, was marked. A metal rod was inserted in the vertical axis of the plant (dotted line) and the distance from the vertical axis (A) to the B point was determined (ab). An upright leaf will show a smaller value than a horizontal leaf. Fifteen plants of each line were field grown from July to September, 1977. The first, third, fifth, seventh, and ninth fully-expanded leaves subtending the curd were measured when the curd was 4-6 cm in diameter. These leaves were samples since these leaves contributed to curd protec- tion. The first fully-expanded leaf was the leaf closest to the curd, and was equal to or longer than 35 cm. 8 .HmucoNfiHo: .mdmv Hmw 2m: .unmfle munmflums .mumv mmw :mz .umoq .momxu amazoumm 03» 059 we knuoeoow mama .H whswfim 11 Figure 2. Longitudinal view of cauliflower plant, showing arrange- ment of leaves. 12 13 Figure 3. Schematic representation of the various leaf types en- countered in cauliflower plants, showing absence of clearly defined angle between leaf base and main axis. l4 VERTICAL | AXIS MAIN STEM Wm 15 Figure 4. Diagram of technique used in measuring leaf geometry in cauliflower. 16 I I A—> ob {—3 I ”I I I I I I errIcAI. I AXIS I I I ' LIA! ROOTS 15 Figure 4. Diagram of technique used in measuring leaf geometry in cauliflower. 16 I l A—> ab {—3 I 900 I I I I I VERTICAL I AXIS I I I I LIA! ROOTS 17 Data from individual plants were entered on IBM cards. Means, variances, and standard deviations were Obtained fOr each population from individual plant data and calculated by the use of the Control Data Corporation 6500 computer. Population means were statistically compared by the use of the 2—tai1ed t-test (Steel and Torrie, 1960). Field Studies Two field experiments, a summer and a fall crop, were conducted during the 1978 growing season. Two inbred parents (6 generations), MSU 831 (P1), and MSU 839 (P2), were used to produce the various populations for the study. Both parents were selections from the cross (Pua Kea x Snowball M) x 'Self—Blanche'. None of these cultivars exhibit upright ‘ F leaf geometry. Families P P F l’ 2’ 1’ 2’ hand pollination in a greenhouse. In both experiments, parents and pro- BCl, and BC2 were produced by genies were grown using a randomized complete-block design with 3 repli- cations. Each replicate contained 15 plants of each parental, backcross, and F1 family and 120 F individuals. The plants were grown in rows 2 spaced 1.07 m apart, with plants spaced at intervals of 0.60 m in the row. Cultural practices were those recommended for commercial production in Michigan. The first experiment (summer crop) was sown April 14 in a cold frame and transplanted in the field June 2. Plants were measured for leaf geometry in July. The second experiment (fall crop) was sown June 14, and transplanted in the field July 7. Plants were measured for leaf geometry in September. Individual plant data were recorded based on the sampling technique developed. Data from individual plants were entered on IBM cards. Means, vari- ances, and standard deviations were obtained for each population from 18 individual plant data and calculated by the use of the Control Data Corporation 6500 computer. Population means were statistically compared by the use of the 2-tailed t-test (Steel and Torrie, 1960). The distribution of the segregating generations was continuous, exhibiting no clear-cut modes which suggested division into phenotypic classes. Consequently, data were analyzed biometrically using the methods outlined by Mather and Jinks (1971). Mather's ABC saclling test (Mather and Jinks, 1971) with A = ZBC— - 1 F1 - P1, B = ZBC2 - F1 - P2, and C = 4F2 - 2F - P1 - P2, was applied to 1 the generation means to determine if the data fulfilled the assumptions of the additive-dominance model. Cavalli's scaling test (Mather and Jinks, 1971) was also conducted. This test utilizes data from all genera- tions to provide estimates of the mean, additive, and dominance effects, derived by the procedure of weighted least squares, using as weights the inverses of the variances of generation means. This joint scaling test also evaluates the goodness of fit of the data to the three-parameter model (mean, additive, and dominance effects), by assuming that the sum of squared deviations weighted with the appropriate coefficients are distributed as a x2 distribution with 3 df. Lack of fit suggests the existence of non-additive gene effects other than dominance. Generation means were also analyzed using the method of Jinks and Jones (1958) to fit a six-parameter model. The symbols used in the six- parameter test and in Cavalli's joint scaling test are those utilized by Gamble (1962). Those parameters are the mean effect, m; the pooled additive effects, a; the pooled dominance effects, d; the pooled addi- tive x additive epistatic effects, an; the pooled additive x dominance epistatic effects, ad; and the pooled dominance x dominance epistatic 19 effects, dd, The equations giving the estimates of gene effects in terms of the generation means are: m = P2 d = -0.SPI'- G.SP; + FI'- 4F;'+ ZBCI'+ ZECE' aa = 2EEI'+ 2§E;'- 4?; ad = -o.s‘;'+ 0.5R;'+ EE;’- RE; dd = R;'+ §;'+ 2?; + 4F;'- 4§EI'- 4§E;' Significance of the various gene effects fOr this model were determined by computing standard errors of the corresponding population means. Heritability in the narrow sense, hzns, was estimated using Warner's procedure (1952) as: 2 h ns - [2vr2 - (VBC1 + vsc2)] / VFZ where VF BC and BC 2’ 2 2’ l’ 2 generations, respectively. A standard error for hzns was derived as the VBCl, and VBC are the variances of the F square root of the following: V(h2ns) = 2{[(VBC1 + VBCZ)2/df F2] + (VBClz/df BCl)-+(VBC22/df BC2)}/ 2 2 . Expected gain from selection (G.S.), was calculated following Allard VF (1960) as G.S. = k Op hzns, where k_is the selection differential, Up is the phenotypic standard deviation of the F and hzns is defined as above. 2, The minimum number of loci differentiating the parents was computed using the methods reported by Castle (1921) and Burton (1951). To determine if the data from the 2 different experiments could be pooled prior to genetic analysis, an analysis Of variance was made using data from both planting times to test time x generation effects. 20 Furthermore, segregating generations from both locations were individual- ly compared using Bartlett's Test of Homogeneity of Variance (Steel and Torrie, 1960). RESULTS AND DISCUSSION Development of Sampling Technique Mean leaf geometry values for the parents MSU 831 and MSU 839 were Obtained by measuring leaves 1, 3, 5, 7, and 9. MSU 831, the open line, with a mean value of 22.37 cm, differed by 9.81 cm from MSU 839, the upright line, which had a mean value of 12.56 cm. Since the sampling of five leaves per plant was time consuming, it was necessary to deter- mine the minimum number of leaf measurements per plant necessary to obtain a valid measure of the plant's leaf geometry. Table 1 shows the means and standard deviations obtained for each leaf and for the various leaf combinations for MSU 839. The 't' and the probability values were ,Obtained by comparing the populations from the various individual leaf measurements or combinations of leaf measurements with the mean of the five leaf measurement (leaves 1, 3, 5, 7, and 9). The means of individual leaf measurements 1, 3, 7, and 9 were each significantly different (5% level) from the mean of the five leaf measurement. The mean of leaf 5 did not significantly differ from the five leaf measure- ment, but had a higher standard deviation. The leaf combination 1 and 5 and 1, 5, and 9 means were not significantly different from the mean of the five leaf measurement. For the MSU 839 parent, the data suggest that sampling leaf numbers 1 and 9 or 1, 5, and 9 would provide a valid estimate of the plant's leaf geometry. 21 22 Table l. Geometric values (cm) for individual leaf and various leaf combinations for MSU 839. Leaf or Mean Leaves (cm) S.D. T Value Prob. l 6.79 3.58 -6.27 .000 3 9.91 3.24 -4.45 .001 5 12.96 4.30 0.52 .612 7 14.84 4.20 2.99 .010 9 18.11 2.80 8.86 .000 1,9 12.54 2.35 0.03 .975 1,5,9 12.68 2.26 -0.48 .641 1.3.5.7.9 12.56 2.30 -- -_ 23 Table 2 shows the results obtained with MSU 831, the horizontal leaf line. As noted in MSU 839, the sampling means of individual leaves 1, 3, 7, and 9 were significantly different (5% level) from the mean of the five leaf measurement. The mean of leaf 5 was not significantly dif- ferent from the five leaf measurement but had a higher standard devia- tion. The mean of leaves 1 and 9 was not significantly different from the five leaf mean. The mean of leaves 1, 5, and 9 was not significantly different from the five leaf mean and had a similar standard deviation, and was judged to be the appropriate sample in this line. On the basis of this study, the mean of leaves 1, 5, and 9 was used to determine leaf geometry. Field Studies — Pooled Analysis The analysis of variance for leaf geometry for the two plantings is presented in Table 3. There was no significance (5% level) for either time or time x generation interaction effects. Table 4 shows the results obtained using Bartlett's test individually comparing the F BCl’ and 2.9 BC variances between the two planting dates. No significant difference 2 (5% level) was noted for either backcross generation. However, a signi- ficant difference (1% level) was noted for the F2 population. Therefore, data were not pooled, and the results Of the two experiments are pre- sented separately. Field Study (Summer) There were no differences (5% level) between the generation means of the reciprocal Fl's or between reciprocal F 's, and therefore, the 2 data were pooled. The frequency distributions for leaf geometry and 24 Table 2. Geometric values (cm) for individual leaf and various leaf combinations for MSU 831. Leaf or Mean Leaves (cm) S.D. T Value Prob. l 17.91 2.82 —6.30 .000 3 20.08 4.40 -2.25 .041 5 20.64 4.24 -l.79 .095 7 24.42 4.12 2.39 .032 9 28.81 1.48 12.81 .000 1,9 23.36 1.91 -l.86 .084 1,5,9 22.45 1.64 -0.24 .812 l,3,5,7,9 22.37 1.59 - -— 25 Table 3. Analysis of variance for leaf geometry for summer and fall crap. Mean Source df Square Prob. Generations 5 89.35 182.69 .00001 Time 1 .10 0.19 .66 Generations x Time 5 .81 1.65 .18 Error 24 .49 26 Table 4. Bartlett's test for homogeneity of variance on F2, BC;, and BC2 generations individually compared between two planting times. Generation X2 P 3C1 1.02 .250 BC2 1.28 .100 F2 14.96 ,001** **Significant at 1% level. 27 graphic frequency distributions are presented in Table 5 and Figure 5, respectively. The F and F2 frequency distributions were slightly skewed toward 1 the upright parent (P2). Sixty-eight and one-half percent of the F2 plants had geometric values lower than the midparent value (16.12 cm). The high frequency of P2 phenotypes in the F2 population suggests that leaf geometry was controlled by a small number of loci in this cross. 0f the total F2 individuals, 3 percent lay above the mean of P1, while 8.9 percent lay below the mean of P2. The means of the different populations (Table 5) substantiate the dominance patterns. The F1 mean was lower than the midparent value (F1 - midparent = -l.46 cm), and was significantly different (5% level) from the midparent value. The F2 showed a similar relationship. The mean of the BC to P2 fell between the F1 and P2 mean, and was closer to the mean of P2 than was the mean of the BC to P1’ which lay between the F and P1 1 means. The parental range was 11.47 cm. The results of Mather's A, B, C and Cavalli's joint scaling tests are shown in Table 6. Factor C was significant in the A, B, C scaling test (1% level), and the joint scaling test was significant (1% level). Significance in either test suggests the possibility of epistasis in the expression of leaf geometry. The suggestion that there may be epistatic gene action made it necessary to abandon the three-parameter model and use Jinks and Jones' six-parameter model to help explain gene action. The E2: dd, and dd_ components of the six-parameter model provide true estimates of the epistatic effects. However, the m, E; and d effects are confounded by epistasis and do not provide valid estimates of the type and magnitude 28 em He.m oe.NH H m m w 03 m «um «a x Ha em Ho.m ee.ea m w m ea e N N aum an x an new em.m me.e~ e m w om om me me me m3 an mfiun an me me.H mo.e~ H o m we we N an mmw am: x Hmw :mz em m~.H wm.o~ e OH ha «a mmm am: we me.H mw.H~ N em mH m H ”a “mm :m: agenda n can: mm mm AN mH AA ma ma HH a aeaaeaueuu uuumfieua mo .02 whoucou mmmfiu .mouo Hoeasm .mmm 3m: x Hmw 3m: mmono Oz» cw AEOV monam> xuuoeoow weed mo mcoflusnfihpmfiv zone:GOHm .m Danae 29 Figure 5. Frequency distributions for leaf geometry (cm) in cauliflower from the cross MSU 831 x MSU 839. Summer. FREQUENCY IS 40 75 25 30 use, I II 13 15 I7 19 21 lEAF GEOMETIY 23 25 31 Table 6. Significance of the A, B, C and joint scaling tests for leaf geometry in a MSU 831 x MSU 839 cross. Summer crop. Test A B C Joint ns ns ** ** **Significant at 1% level. ns, non-significant at 5% level. 32 of gene action independent of epistasis (Ketata et al., 1976). Hayman (1958) proposed that the best approximation to epistasis-free estimates of n, d, and d would be the estimates Obtained from the three-parameter model. Following this suggestion, it appears that additive effects played a major role in the expression Of leaf geometry, although dominance gene action was significant (Table 7). The six-parameter model estimates (Table 7) did not show significant epistatic effEcts (23, dd, and dd). It is possible that the epistatic effects were not of sufficient magnitude to be detected from this test. There is also the possibility that significance was not detected due to a cancelling out of positive and negative gene effects. The dd_epistatic estimate is particularly sensitive to this possible problem (Ketata et al., 1976). Narrow-sense heritability, which reflects the effectiveness of selec- tion for leaf geometry, was estimated at 67 :_26 percent. This estimate may have its limitations, since epistatic effects can inflate the narrow- sense heritability estimate as measured by Warner's method (Ketata et al., 1976). In the scaling test, factor C was significant. However, non- significance of epistatic estimates obtained in the six-parameter model suggests that this narrow-sense heritability estimate is reasonably valid. This estimate, however, should be interpreted as a theoretical maximum, and should be judged as moderate to moderately high. The estimate of genetic advance under selection (G.S.), which evalu- ates the expected rate of genetic gain under directional selection, was 31.5 percent. This suggests that a 31.5 percent increase in the F mean 3 above the F2 mean can be expected if the best 5 percent of the F2 plants are selected. 33 Table 7. Gene effects estimated using three and six-parameter models on generation means in a MSU 831 x MSU 839 cross. Summer crop. Model and Effect Estimates Three-parameter m Sixeparameter a d aa ad dd 15.92**:_-15 5.67**:_.l6 - 1.43**:_,23 11.71 .01 14.65": 0.03 5.34“: 0.42 0.19 :_2.15 1.66 :_2.12 - 0.40 1 0.45 - 0.33 :_7.34 **Significantly different from zero. 34 Two estimates of k, the minimum number of effective factors con- trolling leaf geometry, were computed. Estimates of 1.00 and 1.08 were arrived at using Castle's and Burton's method, respectively. Similar to the narrow-sense heritability estimate, these estimates suggest that a low number of loci are involved in the expression of this trait. The F2 and backcross populations were partitioned into upright and horizontal geometric types based on the arithmetic mean of the two parents (16.12 cm) (Table 8). The following genetic model is proposed: Three major loci (designated 5, B, and 9) make up the parents MSU 831 (P1) and MSU 839 (P2). The proposed genotype of MSU 839 is AAB§Q§_(up- right), and that of MSU 831 is aabbcc (horizontal). The F with a 1. genotype of AaBch, is upright. The observed F2 and backcross ratios suggest a (9:7)(3zl) factorial gene model. A;B;_conditions upright ex- pression regardless of the state of the §_locus. Recessive homozygosity at either the A_or the B_locus in combination with EE.(§E§;EE or A-bbcc) conditions horizontal leaf expression. §;_is epistatic to the expression 0f.33 and db, and conditions upright expression when combined with either gene d_or gene B, The dominance of genes A, d, and §_is incomplete, allowing fer intermediate levels of leaf geometry. The F and backcross 2 generation genotypes A-B-C—, A-B-cc, aaB-C-, and A-bbC- are upright types, while aaB-cc, aabbC-, A-bbcc, and aabbcc are horizontal types (Table 9). The ratios expected (uprightzhorizontal) from the (9:7)(3zl) factorial gene model are 54:10, 1:1, and 1:0 in the F BCI’ and BC2 generations, 2’ respectively. A poor fit (P = <.005) was obtained in both the F2 and BCl generations (Table 8). The aberrant ratio obtained in these genera- tions suggest the presence of modifier genes or an incorrect model. 35 Table 8. Chi-square test for 3—gene model for leaf geometry in the F1, 8C1, and BC2 in the cross MSU 831 x MSU 839. Summer crop. Observed Observed Generation Upright Open Upright Open X2 P F2 207 95 254 47 >10.00 <.005 BC; 9 27 18 18 9.00 <.005 BC2 31 3 34 0 36 Table 9. Proposed genotypes and phenotypes of F2 and backcross.genera- tion individuals. Genotype Phenotype A-B-C- Upright 5:2;22. Upright EE§:§:. Upright d:bb§;_ Upright 33229; Horizontal EEELEE Horizontal 5:2255 Horizontal aabbcc ‘ Horizontal 37 The BC2 generation ratio is not testable using X2, since 0 is the ex- pected frequency in one of the classes. However, based on the assump- tions of the model, the observed ratio of 31 upright:3 horizontal sug- gests agreement with the expectation of the model. Field Study (Fall) There were no differences (5% level) between the generation means of the reciprocal F 's or between reciprocal Fz's, and therefore, the 1 data were pooled. The frequency distributions for leaf geometry and graphic frequency distributions are presented in Table 10 and Figure 6, respectively. The F1 and F2 frequency distributions were slightly skewed toward the upright parent (P2). Approximately 82 percent (81.9) of the F2 plants showed geometric values lower than the midparent value (16.44 cm). The high frequency of P2 phenotypes recovered in the F2 population sug— gests that leaf geometry was controlled by a small number Of loci in this cross. Of the total F2 individuals, 2.0 percent lay above the mean of P1, while 7.4 percent lay below the mean of P2. The means of the different populations (Table 10) substantiate the dominance patterns. The F1 mean was lower than the midparent value (F1 - midparent = -2.20 cm), and was significantly different (5% level) from the midparent value. The F showed a similar relationship. The mean 2 of the BC to P2 lay between the F1 and P2 mean, and was closer to the mean of P2 than was the mean of the BC to P1, which lay between the F and P1 means. The parental range was 10.06 cm. 1 38 NN mo.N eN.NH H e NH Hm mN e «um Na x ad Ne ee.N eo.NH m e NH NN NH a sum an s an moe aN.N mm.mN e EN mN em em mmfi mm Na Na Laue aa ANN om.N MN.eH N Na me mm m an mmw 2m: x Hm” am: oe om.a He.NN m oN oe N «a mmw am: am mN.H ee.HN N on ON 3 e Ha NNN am: nuaeaa u emu: mN NN HN Na Na ma NE NH a eoNueuueuu enamfieua mo .02 whoucou mmmfiu .mono Hamm .mmw 2m: x Hmw 3m: mmouo one ca navy monam> xeuofioom mama mo mcoflusnfiuumfiu zucoscOHm .oH oHnmh 39 Figure 6. Frequency distributions for leaf geometry (cm) in cauliflower from the cross MSU 831 x MSU 839. Fall.! 4O 60 o 1 0 I >92u30u¢u o I IS 50 15 I7 19 21 23 25 lEAF GEOMETRY I3 41 The results of Mather's A, B, C and Cavalli's joint scaling tests are shown in Table 11. Factors A and C were significant in the A, B, C scaling test (1% level), and the joint scaling test was significant (1% level). The results obtained from the three-parameter model suggest that both additive and dominance gene action play a major role in the expres- sion of this trait, with additive effects being of larger magnitude (Table 12). Examining the estimates provided by the six-parameter model, the dd_effects and the dd effects were significantly different from zero. A narrow-sense heritability of 65 1'20 percent was obtained, sug- gesting that selection for this trait should be effective. The limita- tions of this estimate should be considered, due to the presence of epistasis. A genetic advance under selection (G.S.) of 27.8 percent was esti- mated. Values of k_using Castle's and Burton's method were estimated at 1.79 and 1.96, respectively, suggesting that a small number of loci are involved in trait expression. These estimates may also be biased by epistatic effects. The F2 and backcross generations were partitioned into upright and horizontal geometric types based on the arithmetic mean of the two parents (16.44 cm) (Table 13). Based on the ratios expected with the three loci model previously proposed, a good fit was obtained for the F2 (P = .10-.25) and BC1 (P = .25—.50) generations. The BC2 ratio of 76 upright:6 horizontal suggests agreement with the expectation of the model. Compared to the results of the summer crop, a better fit was obtained for the BCI’ and 42 Table 11. Significance of the A, B, C and joint scaling tests for leaf geometry in a MSU 831 x MSU 839 cross. Fall crop. Test A B C Joint **Significant at 1% level. ns, non-significant at 5% level. 43 Table 12. Gene effects estimated using three and six-parameter models on generation means in a MSU 831 x MSU 839 cross. Fall crop. Model and Effect Estimates Three-parameter m 17.08*f: .14 a 5.16*f: .14 d - 3.41**:_,20 x2 64.46 P <.001 Six-parameter m l3.95**: 0.02 a 4.21*f: 0.17 d 1.8] :_1.01 aa 4.01**:_l.00 ad - O.8l**: 0.19 dd - 2.50 + 3.13 **Significantly different from zero. 44 Table 13. Chi-square test for 3-gene model for leaf geometry in the F1, BC;, and BC2 in the cross MSU 831 x MSU 839. Fall crop. Observed Observed Generation Upright Open Upright Open X2 P F; 330 73 340 63 1.88 .10-.25 BC1 28 34 31 31 .58 .25-.50 BC2 76 6 82 0 45 this may be explained by the differential expression of major genes and/or modifier genes under the environmental conditions of each experi- ment . SUMMARY AND CONCLUSIONS A technique for determining leaf geometry in cauliflower was de- veloped. Sampling technique revealed that a valid measure of a plant's leaf geometry was obtained by sampling the first, fifth, and ninth fully-expanded leaves subtending the curd. The progenies of reciprocal crosses between MSU 831 and MSU 839 were evaluated in two field studies (a summer and a fall crOp) to deter- mine the mode of inheritance of leaf geometry. In both plantings, partial dominance of the upright parent was noted. Both additive and dominance gene action played a role in trait expression, with additive action being of larger magnitude. Analysis of generation means suggested that epista- tic gene action played a role in the expression of leaf geometry, with epistasis more pronounced in the fall crop than in the summer crop. Hayman (1958) suggested that the lack of evidence for epistasis in a particular environment may be due to a rise in the error variance rather than to a change in the action of the genes. Analysis of variance of leaf geometry in the summer crop (Table 14) and in the fall crop (Table 15) revealed that the error mean square was higher in the summer crop (12.30) than in the fall crop (3.01). This difference is reflected in the standard errors of the epistatic estimates (Tables 7 and 11), the standard errors being larger in the summer crop than in the fall crop. However, further examination revealed that the magnitude of the epistatic estimates was lower in the summer crop than in the fall crop. Therefore, 46 47 Table 14. Analysis of variance of leaf geometry. Summer crop. Mean Source df Square F Prob. Generations 5 690.57 56.14 .00001 Replications 2 49.93 4.06 .051 Error 10 12.30 48 Table 15. Analysis of variance of leaf geometry. Fall crop. Mean Source df Square F Prob. Generations 5 760.34 252.32 .00001 Replications 2 18.92 6.28 .017 Error 10 3.01 49 it appears that the lack of evidence for epistasis in the summer crop as compared to the fall crop was due mainly to a change in gene action in the different environments rather than to a difference in error. Hayman (1958) also noted that the environment influenced the expression and magnitude of epistasis. Mather and Jinks (1971) reported that the type of epistatic inter- action (complementary or duplicate) can be inferred from the relative signs of d_and dd. Like signs of d_and dd_indicate complementary inter- action, while opposite signs indicate duplicate interaction. In both experiments, d and dd exhibited opposite signs (Tables 7 and 11), sug- gesting duplicate interactions. However, neither the d or dd estimates are significantly different from zero, and the inference as to the type of interaction is probably not valid. Narrow-sense heritability estimates of 67 percent and 65 percent were obtained for the summer and fall crops, respectively. Expected gain from selection estimates of 31.5 percent and 27.8 percent were also calculated. Ratios obatined by partitioning the segregating generations into phenotypic classes based on the arithmetic mean of the parental lines suggested a three loci system (designated 5, B, and C) controlling leaf geometry. A (9:7)(321) factorial of the three loci best explained the observed ratios. Parental line MSU 839 (P2) was designated flfl§§99,and MSU 831 (P1) was designated 222222; The data suggested the presence of modifier genes. Differences in the segregation ratios and in the estimates of the various genetic parameters may be due to the differential expression of the major genes and/or the modifier genes in each of the plantings. 50 Both of the parents used in this study, MSU 831 (P1) and MSU 839 (P2), were derived from the cross (Pua Kea x Snowball M) x 'Self—Blanche'. None of these cultivars exhibit upright leaf geometry. The probable occurrence of uprights from this cross is based on the following observa- tions. (a) No upright progeny were recovered from segregating genera- tions from the cross Pua Kea x Snowball M. (b) Upright phenotypes were recovered in segregating generations derived from the cross (Pua Kea x Snowball M) x 'Self—Blanche' . Based on the proposed genetic model, these observations suggest the pre- sence of an activator loci (D) controlling the expression of leaf geometry. It is proposed that this locus is an activator of gene 9, The following genotypes are proposed for the various cultivars: (a) Pua Kea - aaBBccdd ~ horizontal leaf type (b) Snowball M - aabbCCdd - horizontal leaf type (c) Self-Blanche - AAbbchD - horizontal leaf type (d) Pua Kea x Snowball M - aaBchdd - horizontal leaf type Crosses between Pua Kea and Snowball M will not yield upright progeny, due to the presence of dd_and dd_in both parents. In progeny derived from this cross, the epistatic expression of d_over dd_will not be mani- fested, since the activator locus D_is present as dd, A cross of (Pua Kea x Snowball M) x 'Self-Blanche' will yield upright progeny, since 'Self—Blanche' carries both dd_and 22, Both MSU 831 (P1) and MSU 839 (P2) were selected from this cross. The data suggest that both parents are homozygous DD, Swarup and Chatterjee (1972) concluded that Indian cauliflower ori- ginated mainly but not exclusively from the Cornish cauliflower type of the United Kingdom. Cornish types are characterized by horizontal leaf 51 geometry. Swarup and Chatterjee (1974) later characterized the leaf geometry of 290 inbred lines derived from selfing 200 Indian cauliflower cultivars. A high percentage of inbreds exhibiting upright leaf geometry were obtained. The recovery of a high percentage of upright inbreds that initially originated from horizontal-leaved parents tends to support the proposed epistatic gene model with an activator gene. Crosses between different Cornish types or between Cornish types and other cauliflower types with horizontal leaf geometry could result in recombination of the proposed genes and ultimately may have resulted in the upright phenotypes. CHAPTER II THE INHERITANCE OF PEDICEL WIDTH AND FRUIT DETACHMENT FORCE IN PEPPER (CAPSICUM ANNUUM L.) INTRODUCTION Increase in harvest cost and the unavailability of labor have necessitated the development of mechanical harvesters for many vegetable crops. Recently, research has been directed to the development of a mechanical harvester for the pepper. Consequently, the development of pepper cultivars adapted to mechanized harvest is of importance. Two problems associated with mechanical harvesting of peppers are fruit damage and low yield due to difficulty of fruit removal. The percentage of unharvested and damaged fruit can be minimized by breeding cultivars with low fruit detachment forces. Strains of peppers exist which require little force to separate the calyx from the fruit. The incorporation of this trait into commercial pepper cultivars would be an important step in the development of strains suitable for mechanical harvest. For the fresh market, fruit with pedicels are a definite necessity. The lowering of detachment force of the pedicel from the main stem or minimizing the force necessary to break the pedicel may be realized by developing strains with narrow pedicels. Information regarding the genetic control of pedicel width and fruit detachment force would be helpful in designing an efficient breed— ing scheme for the development of desirable genotypes. The objectives of the present study were: 52 53 To determine the inheritance of fruit-calyx detachment force in mature fruit. To determine the inheritance of pedicel width. To establish relationships of the above characters with fruit length, fruit width, and fruit weight. REVIEW OF LITERATURE Literature on the genetics of pedicel characteristics in pepper is limited. Deshpande (1933) reported that pedicel length was inherited as a quantitative character, with short pedicel dominant over long. Pedicel length in the tomato has also been shown to be quantitatively inherited (Bouwkamp and Honma, 1970). Smith (1951) reported that the 'deciduous' fruit character in pepper was controlled by a single dominant gene. This gene, expressed late in the fruit ripening process, conditions easy separation of the fruit from the calyx. This character was later designated 'soft-flesh', since de- tachment was principally a result of the breakdown of the fruit wall and placental tissue. This character has been of little importance in the development of cultivars adapted to mechanical harvesting. Spasojevic and Webb (1971) reported on the inheritance of abscis- sion of ripe pepper fruit from the calyx, controlled by an incompletely dominant gene. This gene differed from the soft flesh character in that detachment was not a result of fruit and placental breakdown. Stall (1973), using an Ametek Model LG-S strain gauge, determined that detachment characteristics in pepper were under genetic control and found that the pull force of the fruit at the calyx detachment site was significantly correlated to stem-scar diameter. No detailed descrip- tion of the procedure used to measure detachment force was given. A correlation between detachment force and stem-scar diameter has also 54 55 been observed in cucumber (Burnham and Peterson, 1970). Hield et al. (1967) found linear correlations between detachment force and both fruit and stem size in citrus fruits. Villalon and Bryan (1970) evaluated numerous tomato cultivars for ease of fruit-pedicel separation, and feund large differences between cultivars for ease of separation. In- teraction between fruit-pedicel separation fbrce and maturity was signi- ficant. Separation fbrce was found to be related to stem-scar diameter. The results obtained suggested that the same breeding stocks reacted differently with location and season. Inheritance of fruit detachment has been reported in various crops. Barritt (1976) determined that additive gene action controlled calyx removal in strawberry, and obtained a heritability of 84 percent based on parent-offspring regression. Brown and Moore (1975), using a diallel analysis to investigate the inheritance of fruit detachment in straw- berry, determined that both additive and dominance gene action controlled trait expression, and that environmental effects largely influenced de- tachment force. Bassett (1976) investigated the inheritance of pod detachment force in snap beans, and determined that two dominant genes controlled pod detachment force. Narrow-sense heritability was estimated to be 61 percent. MATERIALS AND METHODS' Pedicel Width - Diallel Analysis Six inbred lines, MSU 193, MSU 238, MSU 250, MSU 333, MSU 365, and MSU 393 were used in this study. These lines were selected on the basis of variation in pedicel width and exhibited contrasting fruit charac- teristics. Crosses were made in a greenhouse in all possible combina- tions including reciprocals to produce a 6 x 6 diallel. The Fl's and parents of these crosses were sown in April, 1977 in the greenhouse in peat pots and transplanted to the field in June. A randomized-complete block design with 3 replications was used. Each replicate consisted of one row of each F1 and parent. with each row containing 10 plants. Each row was spaced 1.07 m apart, with plants spaced at intervals of 0.60 m in the row. During the growing season, the plants were irrigated and fertilized as necessary. Data were taken in the following manner: Four mature fruits were removed from each plant. The length, width, weight, and pedicel width were recorded for each fruit. Pedicels were removed from the fruit prior to weighing. Fruit width was measured at the calyx end. Pedicel width was determined with a Helios caliper (VWR Scientific, Chicago, Illinois) by measuring the pedicel immediately below the pedicel-stem attachment zone. The mean of four observations from each plant was used for all calculations. 56 57 Data from individual plants were entered on IBM cards. Means were obtained for each population from individual plant data and calculated by the use of the Control Data Corporation 6500 computer. The genetic analysis was based on the method developed by Jinks (1954) and Hayman (1954a, 1954b). This method allows for the estimation of parameters which provide information about the genetic system con- trolling a quantitative trait. According to this method, variance of the parental and F lines can be partitioned into four parts: 1 variation in the mean effects of the parental lines I93 2; variation which is caused by mean dominance effects of the parents 5; average maternal effects of the parental lines d; variation in reciprocal differences not ascribable to‘g The variation b_is further subdivided into: b1, variance due to directional dominance effects; b2, dominance deviation due to asymmetri- cal gene frequencies; and b3, residual dominance effects. Statistics and the genetic components of variation (D, H H2, and 1’ F) were calculated following Hayman's method (1954a, 1954b). The genetic components of variation measure: D = additive effects of genes H1 = dominance effects of genes H2 = dominance indicating asymmetry of positive and negative gene effects » F = covariation of dominance and additive effects Narrow-sense heritability was estimated using the formula proposed by Crumpacker and Allard (1962): hzns = .250 / (.ZSD + .25H .ZSF + E) 1 where D, H F, and E are the same as those defined by Hayman (1954b). 1, 58 Pedicel Width — F2 Analysis Two inbred parents, MSU 333 (wide pedicel) and MSU 365 (narrow. pedicel) were used to produce the various populations for the study. These parents also exhibited contrasting fruit characteristics (Table 16). Families P1, P2, F1, F2, BCl, and BC2 were produced by hand polli- nation in a greenhouse. Parents and progenies were sown April 13, 1978 in the greenhouse in peat pots and transplanted to the field June 18. A randomized-complete block design with 3 replications was used. Each replicate consisted of one row (15 plants) of each parental, F , and 1 BC generation, and 8 rows (120 plants) of F2 individuals. Each row was spaced 1.07 m apart, with plants spaced at intervals of 0.60 m in the row. During the growing season, the plants were fertilized and irri- gated as necessary. Data were recorded in the following manner: Three mature fruit were removed from each plant. The length, width, weight, and pedicel width were recorded for each fruit. Fruit width was measured at the calyx end. Pedicel width was determined with a Helios caliper (VWR Scientific, Chicago, Illinois) by measuring the pedicel immediately below the pedicel-stem attachment zone. Pedicels were removed from the fruit prior to weighing. The mean of three observations from each plant was used for all calculations. Data from individual plants were entered on IBM cards. Means, variances, and standard deviations were obtained for each population from individual plant data and calculated by the use of Control Data Corporation 6500 computer. Population means were statistically compared by the use of the 2-tailed t-test (Steel and Torrie, 1960). 59 Table 16. Characteristics of the parental lines. Pedicel width investigation. Pedicel Fruit Fruit Fruit Width Length Width Weight. Parent (mm) (cm) (cm) (g) F MSU 333 9.33 9.39 6.27 116.94 MSU 365 2.93 6.35 1.94 6.67 60 The distribution of the segregating generations was continuous, exhibiting no clear-cut modes which suggested division into phenotypic classes. Consequently, data were analyzed biometrically using the methods outlined by Mather and Jinks (1971). Mather's ABC scaling test (Mather and Jinks, 1971) with A = ZBC; - FI'- F}; B = ZBC; - FE'— FE; and C = 2F1'— 51'- FE, was applied to the generation means to determine if the data fulfilled the assumptions of the additive-dominance model. Cavalli's scaling test (Mather and Jinks, 1971) was also conducted. This test utilizes data from all generations to provide estimates of the mean, additive, and dominance effects, derived by the procedure of weighted least squares, using as weights the inverses of the variances of generation means. This joint scaling test also evaluates the goodness of fit of the data to the three- parameter model (mean, additive, and dominance effects), by assuming thattfluesum of squared deviations weighted with the appropriate coeffici- ents are distributed as a X2 distribution with 3 df. Failure of fit sug- gests the existence of non-additive gene effects other than dominance. Generation means were also analyzed using the method of Jinks and Jones (1958) to fit a six-parameter model. The symbols used in the six- parameter test and in Cavalli's joint scaling test are those utilized by Gamble (1962). Those parameters are the mean effect, m; the pooled addi— tive effects, a; the pooled dominance effects, d; the pooled additive x additive epistatic effects, 33; the pooled additive x dominance epistatic effects, ad; and the pooled dominance x dominance epistatic effects, dd. The equations giving the estimates of gene effects in terms of the genera- tion means are: 61 M; d = -o.sFI'- 0.5F;'+ FI'- 4F;'+ 2§EI'+ 2§E;' aa = 2§EI'+ 2§E;'- 4F3' ad = -o.5fi1’+ 0.55;'+ EEI'- EEE' dd = FI'+ §;'+ 2F; + 4F; - 4B6: - 4B6;- Significance of the various gene effects for this model were determined by computing standard errors of the corresponding population means. Heritability in the narrow sense, hzns, was estimated using Warner's procedure (1952) as: 2 _ h ns — [2W2 - (VBC1 + vac2)] / VFZ where VF VBCl, and VBC are the variances of the F 2’ 2 2’ 2 . . 2 . generations, respectlvely. A standard error for h us was derived as the 8C1, and BC square root of the fbllowing: V(h2ns) = 2 {[(VBC1+VBC2)2/df F2] + (VBClz/df BCl)-+(VBC22/df BC2)}/ VFZZ. Expected gain from selection (G.S.), was calculated following Allard (1960) as G.S. = k op hzns, where k_is the selection differential, Up is the phenotypic standard deviation of the F2, and hzns is defined as above. The minimum number of loci differentiating the parents was computed using the methods reported by Castle (1921) and Burton (1951). Detachment Force Two inbred parents, MSU 160 (high detachment force) and MSU 249 (low detachment force) were used to produce the various populations for the study (Table 17). Families P1, P2, F1, and F2 were produced in a greenhouse. Although the BC1 and BC2 families were produced, most of 62 Table 17. Characteristics of the parental lines. Detachment force investigation. Detach Fruit Fruit Fruit Force Width Length Weight Parent (kg) (cm) (cm) (g) MSU 160 4.20 4.04 12.33 35.11 MSU 249 1.77 2.00 6.87 7.57 63 the plants perished in the field. Seeds of the parents and progenies were sown April 13, 1978 in the greenhouse in peat pots and transplanted to the field June 18. A randomized-complete block design with 3 replica- tions was used. Each replicate consisted of 15 plants of each parental and F1 generation, and 120 plants of F2 individuals. Each row was spaced 1.07 m apart, with plants spaced at intervals of 0.60 m in the row. During the growing season, plants were irrigated and fertilized as necessary. Due to poor growing conditions, fruit set was not obtained on all plants. Data were recorded in the following manner: Four mature fruits with the pedicel attached were removed from the plant. A modified Hunter L-lOM spring gauge (Hunter Gauge Co., Lansdale, PA) was used to measure detachment force. The gauge was modified in the following man- ner (Figure 7): A 2.6 cm x 5.0 cm piece of 6 mm pressed steel sheeting was tapped at point A and attached to the threaded extension on the guage. Holes were drilled through the entire width of the steel bar at the end of the bar closest to the gauge, and extended through the bar and into the holes of a common laboratory clamp (Precision Scientific clamp #59-552). The pedicel was then firmly clamped directly above the calyx between the steel bar and the adjustable bar on the clamp (point D). Detachment force was determined by slowly pulling the pedicel per- pendicular to the longitudinal axis of the fruit until the fruit and pedicel separated. Force was recorded in pounds and converted to kilo- grams prior to analysis. The length, width, and weight of each fruit were also measured. The mean of four observations from each plant was used for all calculations. b 4 Figure 7. Instrumentation used to measure fruit detachment force. 65 66 Data from individual plants were entered on IBM cards. Means, variances, and standard deviations were obtained for each population from individual plant data and calculated by the use of the Control Data Corporation 6500 computer. Population means were statistically com- pared by the use of the 2-tai1ed t-test (Steel and Torrie, 1960). RESULTS AND DISCUSSION Pedicel Width - Diallel Analysis Linear regression between the array variances (Vr) and the parent- offspring covariances (Wr), using replicate and reciprocal cross means (Table 18), was conducted to determine if the data fulfilled the assump— tions of the additive-dominance model (Mather and Jinks, 1971). A co- efficient of .57 was obtained. This coefficient was not significantly different from zero (P = .20-.30) or from one (P = .40). A regression coefficient of one is expected if non-additive genetic variation is present as dominance variation. A non-significant coefficient from zero suggests that epistasis may have played a role in the expression of this trait. Since the coefficient was not significantly different from one, it appears that an additive-dominance model may explain gene action for this character. Mean square values from diallel analysis of variance are presented in Table 19. The six error variances (Replication (R) x a, R x b1, R x b2, R x b3, R x g, and R x d) were pooled to give R x £.as a—Eommon error—varianze. A highly significant value for a was obtained, suggest- ing the existence of substantial additive genetic variation controlling the expression of this trait. A highly significant 2 mean square was also obtained, which was mainly effected by the b3 parameter, suggesting the existence of dominance gene action expressed as specific combining 67 68 00.0 00.0 00.0 00.0 00.0 00.0 00002 0000< 00.0 00.0 00.00 00.0 00.0 00.0 00.0 00.0 000 00: 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 000 002 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 000 002 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 000 00: 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 000 00: 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 000 002 0: 0> 000 000 000 000 000 000 000000 =0: :0: =0: :0: =0: 00: .hommom mo 06:00 councw c coozuon mommOHu Hofiamwn mo wow 0 how AEEV £0603 aoowwom :moz .wH oHDMH 69 Table 19. Analysis of variance of diallel table for pedicel width. Mean Source df Square F Ratio 3_ 6 89.73 287.04** b_ 15 255.61 817.69** 21. l 3.22 10.30** b£_ 5 1.33 4.25** El. 9 424.91 1359.28** 3_ S 0.14 0.45 d_ 10 0.37 1.18 Error 70 0.31 **Significant at 1% level. 70 ability in certain hybrid combinations. Reciprocal effects were not detected, as evidenced by the small mean square values for g_and d, Since no reciprocal effects were detected, the mean values of each cross and its reciprocal (Table 19) were used fer the estimation of genetic components of variation D, H H2, and F as defined by Hayman 1, (1954b) (Table 20). Although no estimate of error of these components is available, a comparison of their relative magnitude provides some in- dication of their significance (Mather and Jinks, 1971). The ratio (Bl/D) of 0.33 indicates that both dominant and additive genes contributed to pedicel width. A (“l/D) value equal to one sug- gests complete dominance. The ratio (Hz/4H1) provides an estimate of the average frequency of negative (u) and positive (v) alleles (at loci exhibiting dominance) in the parents. An (Hz/4H1) value of 0.25 would suggest that positive and negative alleles were equally distributed in the parental lines. A value of 0.21 was obtained, suggesting equal allele distribution. The ratio of [4(DH1) + F/DHl — F] was used to estimate the total number of dominant (KD) to recessive (KR) alleles in the parents. A (KD/KR) ratio of 1.13 was obtained, suggesting that more dominant than recessive alleles were present in the parents. Narrow-sense heritability was measured as the ratio of additive variance to total phenotypic variance (Crumpacker and Allard, 1962). An estimate of 84 percent was obatined, providing further evidence of addi- tive gene action in trait expression. Phenotypic correlation coefficients between pedicel width and fruit length, width, and weight for F 's and parents were calculated (Table 21). 1 All correlations were significant at the 1% level, suggesting that 71 Table 20. Components of variation for pedicel width. D H1 H2 F D - H1 H1/1) H2/(4H1) 12.38 4.00 3.35 3.01 8.38 .32 .21 72 Table 21. Correlation coefficients of pedicel width with fruit width, length, and weight. Fruit Fruit Fruit Width Length Weight 0.91 0.54 0.94 All correlations significant at 1% level‘with 1010 d.f. 73 breeding for plants with large fruits and narrow pedicels would be difficult. However, since these correlations were performed on F and 1 parental lines, each of which is genetically unifbrm, the values ob- tained are likely to be dependent on the characteristics of the parental material. Pedicel Width - F2 Analysis There were no differences (1% level) between the generation means of reciprocal Fl's or reciprocal Fz's, and therefore, the data were pooled. The frequency distributions for pedicel width and graphic fre- quency distributions are presented in Table 22 and Figure , respectively. The F1 and F2 frequency distributions were skewed toward the narrow pedicel parent (P2) (Figure 8). 0f the total P individuals, 293 lay on 2 the P2 side of the midparent value (6.13 mm). The low frequency of P1 or P2 phenotypes recovered in the F2 population suggests that pedicel width was controlled by a large number of genes in this cross. No plants of the P1 phenotype were recovered in the F2 population, while only one individual of the P2 phenotypic class was recovered. The means of the various populations are shown in Table 22. The F1 mean fell on the P2 side of the midparent value (Fl - midparent = -1.22 mm), and was significantly different (5% level) from the mid-parent value. The F2 showed a similar relationship. The mean of the BC to P2 fell between the F1 and P2 mean, and was closer to the mean of P2 than was the mean of the BC to P1, which fell between the F1 and P1 means. The range of difference between the parents was 6.40 mm. The results of Mather's A, B, C and Cavalli's joint scaling test are shown in Table 23. All factors were significant at the 1% level in the A, B,(3tests,and the 74 NV mv.o cc.m m Hm N 0um «a x 0m 00 mo.o 00.0 0 m0 on N 0mm 0m x 0m mom 00.o 00.0 N m0 on 000 o0 0 an @000 0a mu m~.o 00.0 mm me 0 0m mom 3m: x mmm :02 N0 00.0 mm.m OH mm «a mom :m: cm ~0.o mm.m 00 o 0m mmm 3m: 00:00m 0 :00: m.m m.w m.n 0.0 m.m m.v m.m m.m cowHMHocou oohwwvom .02 0060:00 00000 .mom 3m: x mmm 3m: 0000a 0:0 :0 AEEV 00:00> £0603 Heofivom mo 0:0003n00u006 xocosconm .NN QHth 75 Figure 8. Frequency distributions for pedicel width (mm) in pepper from the cross MSU 333 x MSU 36S. 30 30 d u FREQUENCY 0- I50 50 2.5 76 ac, ‘ 4.5 6.5 PEDICEL WIDTH A 8.5 77 Table 23. Significance of the A, B, C and joint scaling tests for pedicel width in a MSU 333 x MSU 365 cross. Test A B C Joint ** ** ** ** **Significant at 1% level. 78 joint scaling test was significant (1% level). Significance in either test suggests the possibility of epistasis in the expression of pedicel width. To further understand the possible gene action, Jinks and Jones' six-parameter model was investigated. The E32 ad, and dd components of the six-parameter model provide true estimates of the epistatic effects. However, the m, a, and d_effects are confounded by epistasis and do not provide valid estimates of the type and magnitude of gene action inde- pendent of epistasis (Ketata et al., 1976). Hayman (1958) proposed that the best approximation to epistasis-free estimates of m, a, and d_would be the estimates obtained from the three-parameter model. Therefore, it appears that additive and dominance gene action affected the expression of pedicel width, with additive effects being of larger magnitude (Table 24). Significant (1% level) estimates were obatined using the six- parameter model. Unlike signs of d_and dd_suggests the existence of duplicate gene interactions (Mather and Jinks, 1971). A narrow-sense heritability of 67.5 :_26 percent was obtained, suggesting that selection fbr this trait should be effective. This estimate must be used with caution, since the additive—dominance model did not adequately explain gene action for this character, and epistatic effects can inflate the narrow-sense heritability estimate as measured by Warner's method (Ketata et al., 1976). This estimate should be interpreted as a theoretical maximum, and can be judged as moderate to moderately high. The estimate of genetic advance under selection (G.S.), which evaluates the expected rate of genetic gain under directional selection, was 21.1 percent. A 21.1 percent increase in the F mean above the F mean 3 2 can be expected if the best 5 percent of the F2 plants are selected. 1 79 Table 24. Gene effects estimated using three and six-parameter models on generation means in a MSU 333 x MSU 36S cross. Model and Effect Estimates Three-parameter m 6.23“: .04 a 3.06“: .04 d -1.11**:_ .06 x2 P = <.oos Sixiparameter m 4.70**: .002 a 2.50**: .017 d -2.85**:_.098 aa 0.82**: .092 ad -1.95**:_.013 dd 1.63**:_.305 **Significantly different from zero. 80 Two estimates of k, the minimum number of effective factors con— trolling pedicel width, were computed. Estimates of 12.05 and 12.93 were obtained using Castle's and Burton's method, respectively. These estimates may also be biased by epistatic effects. The correlation coefficients between pedicel width with fruit length, width, and weight were calculated from F generation data (Table 2 25). All correlation coefficients were significant at the 0.01 level, suggesting that selection for plants with large fruit and narrow pedicels may be difficult. The coefficients of determination (r2) corresponding to these correlation coefficient values are 0.11, 0.13, and 0.29, respect- ively (Table 25). The (r2) statistic is equal to the proportion of total .variability of the dependent variable that may be ascribed to the effect of the independent variable. Thus, 11 percent of the variation in fruit width, 13 percent of the variation in fruit length, and 29 percent of the variation in fruit weight can be ascribed to the effeCt of pedicel width. The positive correlation coefficients suggest that, in most cases, selec- tion for fruit with a narrower pedicel will also result inja slight reduction of fruit size. However, the relatively low coefficients of determination suggest that the breeding of individuals with larger fruit and narrower pedicels is,wdthin limits, a feasible objective. Detachment Force There were no differences (5% level) between the generation means of the reciprocal F 's or reciprocal F 's, and therefore, the data were 1 2 pooled. The frequency distributions for detachment force and graphic frequency distributions are presented in Table 26 and Figure 9, respect- ively. 81 Table 25. Correlation coefficients (r) and coefficients of determina- tion (r2) of pedicel width with fruit width, length, and weight in a F2 generation of the cross MSU 333 x MSU 365. Fruit Fruit Fruit Statistic Width Length Weight r 0.33 0.36 0.54 r2 0.11 0.13 0.29 A11 correlations (r) significant at 1% level with 304 d.f. 82 NBH cw.o mv.m v mH NH wH mH mN wH wH m mH HH H «m mHom 0m we om.o vH.m m 0H mm 0 H v 0m mvm 3m: x 00H 3m: 0m hm.o uh.H H v 5 NH m 0m mvm 3m: hm Hm.o om.v N m OH 0 o H m 0; ooH 3m: muszm 0 :00: w.v m.v m.v m.m c.m m.m o.m n.N v.~ H.N w.H m.H N.H :oHuMHocou oouwHwom mo .02 maoucou 0000a .mvm 3m: x ooH 3m: 00000 on» :H wav 063H0> oouow ucoenomuou mo 0:0HuanHu0Hw xocoscoum .0N oHan 83 Figure 9. Frequency distributions for detachment force (kg) in pepper from the cross MSU 160 x MSU 249. 15 d 9' FREQUENCY 20 10 1.2 84 '1 ’1 ’2 I 2.4 3.0 3.6 4.2 DETACHMENT FORCE 4.8 85 The F1 and F2 frequency distributions were skewed slightly toward the high detachment ferce parent (P1). 0f the total F2 individuals, 118 lay on the P1 side of the midparent value (2.98 kg). The F1 mean (Table 26) lay on the P1 side of the midparent (F1 - midparent = -0.16 kg), but was not significantly different (5% level) from the midparent value. The mean and the F mean was significantly different (1% level) from the F 2 1 midparent value. The difference between the parents was 2.42 kg. Two estimates of k, the minimum number of effective factors control- ling detachment fOrce, were computed. Estimates of 1.29 and 1.86 were obatined using Castle's (1921) and Burton's (1951) method, respectively. Due to the lack of BC generation data, no narrow-sense heritability estimate was made. However, the close proximity of the F1 and F2 means to the midparent value suggests a high degree of additive gene action controlling trait expression. This fact, together with the estimate of k_indicating a small number of genes controlling this trait, suggests that selection fbr individuals requiring low detachment force would be possible. Correlation coefficients between detachment force and fruit length, width, and weight were calculated from F data. These coefficients are 2 shown in Table 27. All coefficients were significant at the 1% level. The coefficients of determination (r2) corresponding to these correla- tion coefficient values are 0.44, 0.16, and 0.62, respectively (Table 27). Thus, 44 percent of the variation in fruit width, 16 percent of the variation in fruit length, and 62 percent of the variation in fruit weight can be ascribed to the effect of detachment force. These positive correlation coefficients suggest that, in most cases, selection for fruit with lower detachment force can also reSult in a 86 slight reduction of fruit size. However, the relatively low to interme- diate coefficients of determination suggest that the breeding of indivi- duals with larger fruit and lower detachment forces is, within limits, a feasible objective. 87 Table 27. Correlation coefficients (r) and coefficients of determina- tion (r2) of detachment force with fruit width, length, and weight in the F2 generation of the cross MSU 160 x MSU 249. Fruit Fruit Fruit Statistic Width Length Weight r 0.66 0.40 0.79 r2 0.44 0.16 0.62 All correlations (r) significant at 1% level with 171 d.f. SUMMARY The progenies from 2 mating schemes, diallel analysis and F2 analysis, were evaluated to determine the mode of inheritance of pedicel width, and to determine the relationship of pedicel width with fruit length, width, and weight. Pedicel width was determined to be quantitatively inherited, con- trolled by approximately 12 loci. Both additive and dominance gene action were involved in the expression of pedicel width. Epistatic gene action was also detected. Reciprocal effects were negligible. Narrow-sense heritability estimates of 84 percent and 67.5 percent were obtained from the diallel analysis and F analysis, respectively, sug- 2 gesting that selection for this trait should be effective. Expected gain from selection was estimated at 21.1 percent. In both experiments, correlation coefficients between pedicel width and fruit length, width, and weight were determined. All correlations were positive and signifi- cant at the 1% level, and were of greater magnitude in the diallel analysis. These coefficients suggested that, in most cases, selection for fruit with a narrower pedicel will result in a slight reduction of fruit size. However, the low coefficients of determination suggest that the breeding of individuals with larger fruit and narrower pedicels is possible. Fruit-calyx detachment force was feund to be controlled by a low number of loci, approximately 1 to 2. The expression of fruit-calyx 88 89 detachment force was controlled mainly by additive gene action. Corre- lation coefficients between detachment ferce and fruit length, width, and weight were calculated from F2 data. All coefficients were positive and significant at the 1% level. These coefficients suggested that, in most cases, selection fer fruit with lower detachment force will result in a slight reduction in fruit size. However, the relatively low to intermediate coefficients of determination suggest that the breeding of individuals with larger fruit and lower detachment ferce is possible. The results from this study show that pedicel width and fruit de- tachment ferce are heritable traits. The information obtained on the genetics of these traits and on their relationship to.fruit characteristics is important in the development of cultivars adapted to mechanical harvest. Other plant characteristics, such as branching habit, fruit bearing habit, fruit location, and fruit maturation are factors that are also involved in mechanical harvesting suitability. An understanding of the genetics of these and other components and of their interrelationship to each other will be necessary for the design of optimal breeding schemes for the recovery of desirable genotypes. LITERATURE C ITED 10. 11. 12. 13. LITERATURE CITED Allard, R. W. 1960. Principles of plant breeding. John Wiley and Sons, Inc., New York. 485 pp. Allen, L. H. 1973. Crop micrometeorology: A. Wide-row light penetration. 8. Carbon dioxide enrichment and diffusion. Diss. Abstr. 33:5614. Anderson, M. 1966. Stand structure and light penetration. II. A theoretical analysis. J. Appl. Ecol. 3:41-54. Ariyanayagam, R. P., C. L. Moore and V. R. Carangal. 1974. Selection for leaf angle in maize and its effect on grain yield and other characters. Crop Sci. 14:551-556. Barritt, B. H. 1976. Evaluation of strawberry parent clones for easy calyx removal. J. Amer. Soc. Hort. Sci. 101:569-572. Bassett, M. J. 1976. Inheritance of pod detachment force in snap beans, Phaseolus vulgaris L. HortScience 11:471-472. Bouwkamp. J. C. and S. Honma. 1970. The inheritance of five mor- phological characters in the tomato. J. Hered. 61:19-21. Brown, G. R. and J. N. Moore. 1975. Inheritance of fruit detach- ment in strawberry. J. Amer. Soc. Hort. Sci. 100:569-572. Burnham, M. and C. E. Peterson. 1970. Stem attachment area, a new variable for cucumber breeders. HortScience 5:48-50. Burton, C. W. 1951. Quantitative inheritance in pearl millet (Pennisetum glaucum). Agron. J. 43:409-417. Castle, W. E. 1921. An improved method of estimating the number of genetic factors concerned in cases of blending inheritance. Science 54:223. Chang, T. T. and 0. Tagumpay. 1970. Genotypic association between grain yield and six agronomic traits in a cross between rice varieties of contrasting plant type. Euphytica 19:356-363. Crumpacker, D. W. and R. W. Allard. 1962. A diallel cross analy- sis of heading date in wheat. Hilgardia 32:275-318. 90 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 91 Deshpande, R. B. 1933. Studies in Indian chillies. 3. The in- heritance of some characters in Capsicum annuum L. Indian J. Agr. Sci. 3:219-300. Donald, C. M. 1968a. The breeding of crop ideotypes. Euphytica 17:385-403. Donald, C. M. 1968b. The design of a wheat ideotype. Proc. Third Int. Wheat Genetics Symposium: 359-369. Australian Academy of Science, Canberra. Duncan, W. G. 1971. Leaf angles, leaf area, and canopy photo- synthesis. Crop Sci. 11:482-485. Fralich, W. G., W. G. Pollmer and D. Klein. 1977. Performance of upright-leaved liguleless-Z and liguleless-3 maize hybrids adapted to central European climatic conditions. Z. Pflanzenzuchtg. 72:134-144. Gamble, E. E. 1962. Gene effects in corn (gga_mays L.). 1. Separation and relative importance of gene effects for yield. Can. J. Plant Sci. 42:339-348. Gardener, C. J., J. W. Tanner, N. C. Stoskopf and E. R. Reinbergs. 1964. A physiological basi52fimrthe yield performance of high and low yielding barley varieties. Agron. Abstracts.' Annual' Meetings of the American Society of Agronomy, Kansas City. Hayman, B. I. 19543. The analysis of variance of diallel tables. Biometrics 10:235-244. Hayman, B. I. 1954b. The theory and analysis of diallel crosses. Genetics 39:789-809. Hayman, B. I. 1958. The separation of epistatic from additive and dominance variation in generation means. Heredity 12:371-390. Hicks, D. R. and R. E. Stucker. 1972. Inheritance of yield com- ponents and morpho-physiological traits in barley, Hordeum vulgare L. Crop Sci. 12:283-286. Hield, H. Z., L. N. Lewis and R. L. Palmer. 1967. Fruit-stem detachment forces. The California Citrograph. 52:420-424. Hopper, N. W. 1970. The evaluation of upright leaves and the use of sucrose and glutamate in increasing the yield of corn. Ph.D. Thesis, Iowa State Univ., Ames, Iowa, USA. Jinks, J. L. 1954. The analysis of continuous variation in a diallel cross of Nicotiana rustica varieties. Genetics 39:767- 788. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 92 Jinks, J. L. and R. M. Jones. 1958. Estimation of the components of heterosis. Genetics 42:223-234. Ketata, H., L. H. Edwards and E. L. Smith. 1976. Inheritance of eight agronomic characters in a winter wheat cross. Crop Sci. 16:19-22. Kuroiwa, S. 1970. Total phytosynthesis of a foliage in relation to inclination of leaves. In; [Prediction and measurement of photosynthetic productivity. Proc. IBP/PP Techn. Meet., Trebon 1969, 79-78. Pudoc, Wageningen, Netherlands. Mason, L. and M. S. Zuber. 1976. Diallel analysis of maize for leaf angle, leaf area, yield, and yield components. Crop Sci. 16:693-696. Mather, K. and J. L. Jinks. 1971. Biometrical Genetics. Cornell Univ. Press, Ithaca, New York. 382 pp. Mock, J. J. and R. B. Pearce. 1975. An ideotype of maize. Euphytica 24:613-623. Nigam, S. N. and J. P. Srivastava. 1976. Inheritance of leaf angle in Triticum aestivum L. Euphytica 24:457-461. Pearce, R. B., R. H. Brown and R. E. Blaser. 1967. Photosynthesis in plant communities as influenced by leaf angle. Crop Sci. 7:321-324. Pendleton, J. W., G. E. Smith, S. R. Winter and T. J. Johnston. 1968. Field investigations of the relationship of leaf angle in corn (Ega_may§ L.) to grain yield and apparent photosynthesis. Crop Sci. 60:422-424. Russell, W. A. 1972. Effect of leaf angle on hybrid performance in maize (Zea mays L.). Crop Sci. 12:90-92. Rutger, J. N. and L. V. Crowder. 1967. Effect of high plant density on silage and grain yield of six corn hybrids. Crop Sci. 7:182-184. Sakamota, C. M. and R. H. Shaw. 1967. Light distribution in field soybean canopies. Agron» J. 59:7-9. Seetharaman, R. and D. P. Srivastava. 1971. Inheritance of height and other characteristics in a rice mutant. Indian J. Genet. Plant Brdg. 31:237-242. Sinclair, T. R. 1972. An evaluation of leaf angle effects on maize photosynthesis and productivity. Diss. Abstr. 32:6768. Smith, P. G. 1951. Deciduous ripe fruit character in peppers. Proc. Amer. Soc. Hort. Sci. 57:343-344. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. S3. 93 Spasojevic, V. and R. E. Webb. 1971. Inheritance of the abscis- sion of ripe pepper from its calyx. (In Hungarian, English summary). Arhiv. Bioloskih. Nauka 23:115-119. Stall, W. M. 1973. An evaluation of fruit detachment characters in pepper (Capsicum annuum L.). Ph.D. Thesis, Univ. of Florida, Gainesville, Florida, USA. Steel, R. G. D. and J. H. Torrie. 1960. Principles and procedures of statistics. McGraw-Hill, New York. 481 pp. Swarup, V. and S. Chatterjee. 1972. Origin and genetic improve- ment of Indian cauliflower. Econ. Bot. 26:381-393. Swarup, V. and S. Chatterjee. 1974. Improvement of Indian cauliflower. Indian J. Genet. 34:1300-1304. Tanner, J. W., C. J. Gardener, N. C. Stoskopf and E. Reinbergs. 1966. Some observations on upright-leaved type small grains. Can. J. Plant Sci. 46:690. Villalon, B. and H. H. Bryan. 1970. Evaluation of fruit-pedicel separation of fresh-market tomato varieties for mechanical harvest. Proc. Fla. State Hort. Soc. 83:127-130. Warner, J. N. 1952. A method for estimating heritability. Agron. J. 44:427-430. Whigham, D. K. and D. G. Woolley. 1974. Effect of leaf orienta- tion, leaf area, and plant densities on corn production. Agron. J. 66:482-486. Winter, S. R. and A. J. Ohlrogge. 1973. Leaf angle, leaf area, and corn (EE§.E§Z§.L°) yield. Cr0p Sci. 65:395-397. Yap, T. C. and B. L. Harvey. 1972. Inheritance of yield components and morpho-physiological traits in barley, Hordeum vulgare L. Crop Sci. 12:283-286.