j GENETIC ANALYSIS or MORPHOLOGICALI . * ; CHARACTERISTICS OF FIELD BEAN : ' V. g; (PHASEOLUS VULGARIS L. ) As. EXPRESSED IN A-- ~ , - , ’ I DIALLEL CROSS -. - gf ‘ "Dissertationforthe"Dégreeo,f'Phy.D.A f g; MICHIGAN STATE UNIVERSITY * ., : , THONGCHAIIONGUTHNSRI' - .' ”A. I ., V 7' - '. z ' -‘ ‘ , , ._ .. .. ' . , .. I -. r‘ I. ~ ' ‘ -. .I. I n . I .i ' , ,. . . .. ,. - - .. . . . ,. . V . ' ' I. .A . Illlllllllllllllll‘lllllllllllllIlllUlllllllllllllllll . ‘ many 3 1293 10564 1132 Michigan Stab Univuity This is to certify that the thesis entitled GENETIC ANALYSIS OF MORPHOLOGICAL CHARACTERISTICS OF FIELD BEANS (Phaseolus vulgaris L.) AS EXPRESSED IN A DIALLEL CROSS presented by Thongchai Tonguthaisri has been accepted towards fulfillment of the requirements for Ph.D. degree m Department of Crop and Soil Sciences Major professor Date W b ”g / 9) 75 0-7639 ABSTRACT GENETIC ANALYSIS OF MORPHOLOGICAL CHARACTERISTICS OF FIELD BEAN (PHASEOLUS VULGARIS L.) AS EXPRESSEDEIN A DIALLEL CROSS BY Thongchai Tonguthaisri The increased yield in rice and wheat has been achieved in conjunction with, and to a large extent as a consequence of, changes in plant type. Both morphological and certain physiological characteristics of these plant types are strongly associated with grain yields. With proper management during the growing season these new plant types respond positively toward high yields. In field bean, however, ideal plant types for dif- ferent ecological zones are only beginning to be identi- fied. Perhaps as the traits contributing to high yield in field beans are more clearly identified and better understood genetically, it will be possible in beans, as in rice and wheat, to set about in a methodological fashion the creation of new higher yielding types. The focus of this thesis is upon the genetic analysis of morphological characteristics believed to be associated with improved yield potential. Thongchai Tonguthaisri Inheritance of morphological characteristics of field beans (Phaseolus vulgaris L.) was studied in 5x5 and 8x8 diallels in 1973 and 1974. Length of pod in the 8x8, F and 5x5, F2, number 1 of seeds per pod (5x5, F2) and loo-seed weight were found to be inherited additively. Additivity was not detected for days to first flowering, duration of flowering, days to maturity, total plant weight, number of main stem branches, number of main stem nodes, total number of nodes, length of internode (8x8, F and 5x5, F2) number 1 of racemes (8x8, F1 and 5x5, F2), number of seeds per pod (5x5 and 8x8, F1), number of pods per plant, pod dry weight per plant, number of seeds per plant and seed dry weight per plant. Heterosis was generally high in the F1 generation for most traits. The level of heterosis decreased in many traits in the F2 generations due to the increase of homozygosity. The heritability estimate of loo-seed weight was the highest (82%) in the 5x5, F data. Length of pod in 1 the 8x8, F had a heritability estimate of approximately 1 100% and loo-seed weight was 90% in the same set of data. However, in the 5x5, F2 data, number of seeds per pod and number of seeds per plant were both 90% whereas lOO-seed weight was 81%. The rest of the traits studied had much lower heritability estimates (less than 30%). Thongchai Tonguthaisri The variety Swedish Brown (SB) had the highest harvest index in both the 5x5, F and 8x8, F data. The l 1 strain 0674 had the lowest H.I. in the 5x5, F1 and strain 0685 had the lowest in the 8x8, F In the F2, the 1' highest H.I. values were very close; Black Turtle Soup (BTS), and Seafarer (SEA) were the highest with the value of .61 whereas SB and strain 72-7427 were the next highest with the value of .60. The cross 72-7427 x Jules gave the highest H.I. value of .71 in the 8x8, F1; the cross SB x 72-7427 gave the highest H.I. of .61 in the F set of data. 2 Harvest index was not found to be directly related to yield nor of high heritability but it can be useful as one important measure of efficiency of the plant. Black Turtle Soup was found to be a promising variety in yielding ability. It has highest mean values of number of pods per plant, pods weight per plant, number of seeds per pod and number of seeds per plant. Being a black-seeded variety seemed to associate with high yield. However, if highly mechanized planting is concerned, Tui may be a good substitution. This is because of the erect plant type of Tui which can be grown at a higher number of plants per unit area. Besides Tui is only slightly inferior to BTS in seed yield. With higher number of plants per unit area Tui may be equal to or better than BTS in yield. Thongchai Tonguthaisri The evaluation of SBA and strain 0674 may not be completely accurate in some characteristics because of their sensitivity to ozone injury. Each variety seemed to have the highest mean value of a particular trait. It, therefore, depends on the breeder to exploit these par- ental lines and incorporate those traits which have addi- tive effect into a new variety. GENETIC ANALYSIS OF MORPHOLOGICAL CHARACTERISTICS OF FIELD BEAN (PHASEOLUS VULGARIS L.) AS EXPRESSED IN A DIALLEL CROSS BY Thongchai Tonguthaisri 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 1976 To my parents, To M. R. Kukrit Pramoj ii ACKNOWLEDGMENTS The author wishes to express his sincere appreci- ation to Dr. M. W. Adams for his valuable guidance, patience, and understanding that made this graduate study possible and rewarding. The author also is indebted to his critical review of this manuscript. A grateful acknowledgment is extended to the mem- bers of the guidance committee: Dr. S. Honma, Dr. J. E. Grafius, and Dr. L. Copeland. He wishes to thank Dr. D. H. Smith who served in the examining committee. The author thanks Dr. J. P. Bennett for computer assistance, Dr. C. Cress for statistical advice, Mr. J. Taylor for field assistance, and Dr. J. Wiersma for sug- gestions and encouragement. Special thanks are also extended to several Thai friends who kindly assisted in seeds preparation, field planting in 1974. He also thanks friends who assisted in preparation of the figures for this manuscript. A grateful acknowledgment is extended to the Rockefeller Foundation for financial support, without which his graduate studies could have never continued. iii Finally, he deeply appreciates the moral support from.his wife, Tatsanai. Her dedication and understanding were most valuable. iv INTRODUCTION TABLE OF CONTENTS 0 O O 0 REVIEW OF LITERATURE . MATERIALS AND METHODS . Missing Data . . . Diallel Analysis. . Test of the Significance Line and the Intercept Heritability Estimate . RESULTS AND INTERPRETATION Days to First Flower 5 x 5, 8 x 8, 5 x 5, Duration of Flowering 5 x 5, F F: o o o F 2 o o 0 F1 0 o 0 Days to Maturity. . F1 . . . F O O . F% O O 0 weight. . F O O O F; O O 0 F2 0 o o the Main Stem Branches. . Fl 0 o 0 Regression Page 10 14 15 21 21 22 23 23 37 43 43 45 45 52 54 54 59 61 61 Number of Main Stem Nodes 5X5, F1 0 o o 0 Total Number of Nodes . 8X8'F1 o o o 0 5X5, F2 0 o o 0 Length of Internode . . 8x8,F1. . 5X5,F2 o o o 0 Number of Racemes Per Plant 8X8, Fl 0 o o 0 5x5, F2 0 o o 0 Number of Pods Per Plant 5 X 5’ F1 0 o o o 8 X 8' Fl 0 o o o 5 X 5' F2 0 o o 0 Length of Pod . . . . 8X8, F1 o o o o SXS'FZ o o o o Pod Dry Weight Per Plant 5' F1 0 o o 0 F1 0 o o o 5, F2 . . . . unmtn x> Hl when the intercept is positive and it is then that the dominance is partial or incomplete. When D = H1, the line passes through the origin, and dominance is complete. If D < H then the l intercept is negative and over dominance is indicated. 3. The position of the array point nearest the origin indicates that a parent contains a preponderance of dominant genes and when furthest from the origin indi- cates that the parent contains fewer dominant genes or mostly recessive genes. If there is no dominance, H = 0 l and all the array points cluster at a single point where: Vr 1/4D l/4Vp Wr = l/2D l/2Vp 21 Test of the Significance of the Regression Line and the Intercept a) To test whether the regression line is sig- nificantly different from either unity or zero, the following t-test was used: b - bo t=_—s_— b where: b = regression coefficient sb = standard deviation of the regression coefficient bo = l or 0 respectively b) To test whether the intercept is significantly above or below the origin. t = I/sy where: y = bo + bl X1 bo = a = intercept bl = regression coefficient 2 s§=sz(l_+(x-i’)2=sz(l+g) n 2x n 2x2 Heritability Estimate The heritability estimate in this program followed that used by Crumpacker and Allard (1962), i.e. 2 1/4D h = 1/4D + 1/4Hl - 1/4F + E RESULTS AND INTERPRETATION The summary of mean squares and the significance of variance ratios from the analysis of variance of the 5 x 5, parental and F hybrids in 1973 are shown in 1 Table 2. Significant differences occurred for the 14 traits studied among the 15 populations or genotypes com- pared (5 parents and progenies from 10 crosses). However, number of main stem nodes per plant did not show signifi- cant differences among the genotypes. It was assumed that the prerequisites for further analysis were ful- filled. The results for each trait will be presented and discussed proceeding from the 5 x 5 diallel F in 1973 1 to the 8 x 8 diallel F and 5 x 5, F2 in the 1974 trial. 1 In the 1974 trial, there were 14 parental means in each replicate. Since there were only 6 F2 means from each cross, six parental means were taken randomly out of the 14 means and averaged for the purpose of comparison. On the other hand, since there were only 2 Fl means from each cross, 2 of the parental means were than taken randomly from among the 14 means and averaged to compare with the F1 means. It can therefore be noticed from 22 23 Tables 3 and 5 that the same parents do not have the same mean values even though they were planted in the same year. For example, the mean value of days to flowering of BTS in Table 5 is 44.9 whereas the mean value in Table 3 is 45.2 and so on. Therefore, the F2 families will be compared to their parental means in the same table and the F hybrids of the 8 x 8 will be compared to their 1 parental means as presented in Table 3. Days to First Flower The number of days from emergence to first day the plant flowered was recorded on a single plant basis. SXS’Fl The ranking of the mean values of the 5 parents in the order of first to last flowering was: strain 72-7427 25.3, SB = 28.7, SEA = 30.5, strain 0674 = 35.1 and BTS = 36.9 (Table 1). Strain 72.7427 is a red kidney line with large leaves and determinate plant type. SB has approxi- mately the same plant type as strain 72-7427 and also has several morphological characters in common. SBA and strain 0674 represent the navy bean determinate type with small leaves. They also possess several additional char— acters in common. BTS is the only indeterminate type, with small leaves and purple flower, in this parental group. As shown in Table l, BTS flowered latest among the 5 parents. It has been generally observed in beans 24 0.00 0.0 H.0H 0.00H 0.00 N.5~ N.0N <00 x H.00 H.0 0.HH 0.00 0.50 0.0m 5.00 0500 x 0.0m 0.0 0.5 0.H0 0.50 0.~N 0.5m mm N 0.0a 0.0 0.5 0.00 0.05 n.0H 0.0m mam x5N05IN5 0.00 0.0 H.0H 0.00H 0.H0 0.0m 0.0m 5N05I~5 x 0.00 0.0 0.0 0.00 0.00 ~.0H 5.00 0500 x m4~0 0.0 0.0 0.0NH 0.00 0.0m H.Hm 00 x 0.05 0.5 0.0 0.00H 0.00 0.0m 0.00 080 x <00 0.00 ~.0 0.0a ~.0HH 0.00 0.m~ 0.5m 5~05IN5 N H.00 0.0 H.0 N.00 0.~0 H.0H 0.0m <00 x 0.00 0.0 H.~H 04HOH 0.00 0.H0 0.50 mm x 034 .0.» Tm .022 THE Dow T3 95 x 38 H.H0 H.0 0.0a 0.00H 0.00 H.5N 0.5N 5N05IN5 x «.00 H.0 0.0a 0.0NH 5.00 0.0m 0.0N <00 x 0 00 0.0 0.NH 0.00H 0.00H 5.0m N.5m 0500 x 0.50 0.0 H.5 5.0HH 0.H0 0.0m 0.00 090 x 00 0.00 0.0 0.0 0.00H 0.~0 0.0m 0.00 5N05I~5 x H.00 5.5 N.0 0.00H 0.H0~ 0.~N 0.0m <00 x ~.0~H 5.0 0.5 0.~0N 0.00H 0.0m 0.00 0500 x 0.00 0.0 0.5 0.00 0.00 0.~N .0.00 mm x 090 0.0a H.0 0.0 0.05 0.00 5.0m m.0~ 5N05IN5 ~.00 0.0 0.0 0.00 0.05 5.NN 0.0m cme .H magma 25 H0. N.m0 H.HON N.00 0.0 0.N0 umm \.u3 %HQ flmmm udem\m~uwmm x» .u3 fimmmlOOH v0m\mvmmm * \.u3 .AHQ flom A.udoov .H mabme 26 moo. Na Auouumv sum x mum «4mmm.~ moo. 0N Ammssuoamuv mmfisfiamm mmo.N coo. m maofiuwofiaamm m> .m.z xmvcH umo>umm m0o.~0m Ham.w «km. mau.m~m N5 Auouumv awn x gum «000N.N 0N~.050~ *«H00.05 00H.H00 *«0H0.0 00H.H ««~00.0 0H0.000N 0m Ammahuosmuv mowawamm 00H. 000.00H 050. 000. 000. ~0N. ~0H. H00.00H 0 msoaumowammm m> , .m.z m> .m.x m> .m.z m> .m.z 500 szuamam\.us sun nmmm AN0 .03 cmmm.ooa wom\mommm mo 0 Axe “amflm\muom mo 0 000.0000N 000.550H 000.0H 000.0 N5 Auouumv awn x 000 «*0NH.0 H05.H0050 «*00H.~ 000.0H00 ~00.H 5N0.00 ««050.N 500.0 0N AmeMuocmuv mmHHstm 00H. H~0.0000 05H. 050.000 HON.H H05.0~ H00. 000.H 0 mcowumoaammm m> .m.z m> .m.z m> .m.z m> .m.z mmwoz mwr—UCNHM unmam\wvmom mo 0 usm0m\.u3 mun com” amum can: wo 0 amum cam: mo 0 000.H000 05H.H0 000.0 005.0 N5 Auouumv 3mm x 000 «x000.~ 550.5000 «x500.0 055.50N «x000.0 000.00 ««0HH.0H 000.00H 0N Amwmuuosouv mmaaaamm ~00.H 0H0.00H0 000. 0H0.0N 000.H 000.0H ~05. 000.0 0 acofiumowammm m> .m.z m> .m.z 0> .m.z m> .m.z wcwumaoam .u3 mun ucmam huausumz ou mama mo sowumuam uoaoaw uma Ou mama .m.0 .moaumu mosmauw> mo cosmowwacwam 0am mmumsvm came 00 mumaasm .050H .mcwunzn Hm 0am Handmade .0 x 0 mag 00 mosmwum> mo mamaaaa< .N «Hsme 27 that the indeterminate type tends to flower later and over a somewhat longer period than the determinate plant type. (Early and late flowering genotypes do, however, occur in both types.) Bliss (1971) reported that the determinate plant type in segregating generations of crosses of 7 cultivars flowered earlier than the indeter- minate type. From the mean values of the F1 hybrids in Table 1, it appears that 8 out of the 10 hybrids had mean values greater than that of the mid-parent value; four hybrids had mean values greater than that of the late-flowering parent. Two hybrids had mean values smaller than the mid- parent value. The Wr/Vr graph (Figure 1) shows that the regres- sion of Wr on Vr is significantly different from zero but not significantly different from unity (b = .62 i .19). The regression line intersects the Wr axis above the origin (a = 3.12 i 4.15) but not significantly different from zero, indicating partial to complete dominance. The /'Hl/D value of 1.76 indicates over dominance. On balance the evidence suggests partial to complete domr inance. Strain 0674 appears to contain a preponderance of dominant genes for later flowering. It is believed that ozone injury caused the delay in growth and develop- ment of strain 0674. The later development when climatic conditions were favorable resulted in long duration of 28 .H so mud o mama . musmwm 05. u 0 u 0 Mom u m am 0 .m u a .3 00.0 n a u Q How u 0:6 00.0 n 0 n a How u« .> 00 on ON 0. I q o 0.. In." mwtvu... «0.0 u a « 0..“ N0." 0 . Or .0 0)) no N.‘ hflQhuNh u m. .1 ON (mm u G. vhmo u 0. Q mm H. N. aka 0 w. . On 29 flowering of this genotype. Parent 4 (SEA) contains a preponderance of recessive genes for earlier flowering. Parents 2 (SB) and 5 (72-7427), which are very similar in several phenotypic characters, seem to have similarly behaving genotypes with respect to this trait. Parent 1 (BTS), which is distinctively different from the other 4 parents, appears to be an outlier in this case. Since BTS, SB and strain 72-7427 are located in the middle of the regression field it is indicative that they contain an intermediate balance of recessive and dominant genes for appearance of first flower. The heritability was 29% for this trait, as estimated in this Fl diallel set. The mean number of days to first flower of the 8 parents, from early to late, ranked in the following order: strain 72-7427 = 28.0, SB = 33.8, SEA = 34.8, strain 0674 = 35.4, Jules = 36.1, BTS = 45.2, strain 0685 = 48.6 and Tui = 50.5 (Table 3). The analysis of variance showed that there were no significant differences among replicates, but highly significant differences among the genotypes existed (Table 4). All of the original 5 varieties used in the 1974 trial flowered later than in 1973. Different environmental conditions in the two years presumably caused these differences. It was observed that there was a drought period in 1974 at 3() 0.00 0.05 0.00 0.000 0.05 0.000 5.50 0.00 00000 x 0.00 0.05 0.00 0.000 0.00 0.000 0.00 0.00 000 x ~.H¢ m.mm m.m~ n.000 0.0m o.Nm w.mm 0.0m mmoo x 0.000 0.00 0.00 0.000 0.50 0.000 0.00 0.00 5005I05 x 0.00 0.00 0.00 0.000 0.00 0.05 0.00 0.00 0500 x 0.000 0.00 0.50 5.550 0.00 0.000 5.00 0.00 00 x 0.00 0.00 0.00 0.000 0.000 0.000 0.000 0.00 090 x <00 0.00 5.00 0.00 0.000 0.50 0.000 0.00 0.00 00000 x 0.00 0.05 0.00 5.000 5.00 0.000 0.000 0.00 009 x 0.00 0.05 0.00 0.000 0.00 0.500 5.50 5.00 0000 x 00.000 0.00 0.00 0.000 0.00 5.00 0.00 0.00 5005:05 x 0.00 0.00 0.00 0.000 0.00 0.00 0.05 0.00 <00 x 0.00 0.00 0.00 0.000 0.50 0.000 0.00 0.00 00 x 0.00 5.00 0.00 5.500 0.000 0.000 0.000 0.00 000 x 0500 0.000 0.00 0.00 5.000 0.00 5.500 0.00 0.00 00000 x 0.000 5.00 0.00 0.000 0.00 0.000 0.000 0.00 009 N 0.000 0.00 0.00 0.000 0.05 0.000 0.000 0.00 0000 a 0.000 0.00 0.00 5.000 0.00 0.00 0.50 0.00 5005105 N 0.000 0.00 0.00 0.550 0.00 0.000 0.00 0.00 <00 x 5.00 0.00 0.00 0.000 0.00 0.000 0.00 0.00 0500 x 0.000 0.00 0.00 0.000 0.00 0.000 0.00 090 x 00 0.000 0.00 0.00 5.000 0.00 0.000 0.00 0.00 00000 x 0.50 0.00 0.00 0.00 0.05 0.000 0.000 0.50 008 x 5.000 0.000 0.00 0.500 0.000 0.000 0.500 0.00 0000 x 0.000 0.00 0.00 0.000 0.00 0.500 0.00 0.00 5005105 x 0.00 0.05 5.50 0.000 0.00 0.000 0.00 0.50 <00 x 5.00 0.00 0.00 0.000 0.000 0.000 0.500 0.00 0500 N 5.000 0.00 0.00 0.000 0.00 0.00 0.00 0.00 00 x 080 0.00 0.00 0.00 0.000 0.00 0.00 0.00 0.00 <00 0.00 0.00 0.00 0.000 0.00 0.00 5.00 0.00 0500 0.000 0.00 0.00 0.000 0.00 0.00 0.00 0.00 00 0.00 0.00 0.00 0.00 0.50 5.00 0.00 0.00 000 5550 com 00 :0: uam0m uam0m AsavwvocumHQH movos 00 0600 .us >u0u=umz H0300.0 numsoa ~0mu0>< \mvom 0 \moamumx 0 we sawsma nons::.uoe 000 uam0m On @000 un0 cu man .0500 C0 mommouo 0000m00 0 x 0 wsu mo m00unm£ 00 can mquumn ~00 mo=0m> sum: .0 00009 31 00. 0.00 0.000 5.00 0.0 0.000 00000 x 00. 0.50 0.000 0.00 0.0 0.000 000 x 00. 0.00 0.000 0.00 0.0 0.00 0000 x 00. 0.00 0.000 0.00 0.0 0.00 5005105 x 00. 0.00 0.000 5.00 0.0 5.00 0500 x 50. 0.05 0.000 0.00 0.0 0.00 00 x 00. 0.00 0.000 5.50 0.0 0.000 090 x <00 00. 0.05 0.000 0.00 0.0 0.00 00000 x 00. 0.00 5.000 0.00 0.0 0.000 009 x 00. 0.00 0.000 0.50 0.0 0.00 0000 x 00. 0.00 0.000 0.00 0.0 0.00 5005105 x 00. 0.00 0.000 0.00 0.0 0.50 <00 x 00. 0.00 0.000 0.00 0.0 0.00 00 x 00. 0.00 0.050 0.00 0.0 0.000 090 x 0500 05. 0.05 0.000 0.00 0.0 0.00 00000 x 00. 0.05 0.000 0.00 5.0 0.00 009 x 00. 0.05 0.000 0.50 0.0 0.00 0000 x 00. 0.00 0.000 0.00 5.0 0.00 5005105 x 00. 0.00 0.000 0.00 0.0 0.000 <00 x 00. 0.00 0.500 0.00 0.0 0.05 0500 x 00. 0.00 0.000 0.50 0.0 0.00 000 x 00 00. 0.00 0.500 5.00 0.0 0.000 00000 x 00. 0.05 0.000 0.50 0.0 0.00 000 x 00. 0.000 0.000 0.00 5.0 0.000 0000 x 00. 0.00 0.000 0.00 0.0 0.000 5005105 x 00. 0.05 0.050 0.00 0.0 0.000 <00 x 00. 0.05 0.000 0.00 0.0 0.000 0500 x 00. 0.00 0.000 0.00 5.0 0.05 00 x 000 00. 0.00 0.000 0.00 0.0 0.00 <00 00. 0.00 0.000 5.00 0.0 0.00 0500 00. 0.00 0.000 5.00 0.0 0.00 00 00. 5.00 0.000 0.00 0.0 0.05 000 unmmvcH AEWV “GUM“ “HHQHQ Aawv :N: 2%.. “av UCNHW umm>hm0 \.u3 000 0000 \mvwom 0 .u3 vwom1000 vom\mvmmm 0 \.u3 hum vom 0.»:000 .0 m0an J32 0.000 0.00 0.00 0.50 0.50 0.000 0.00 0.00 0:0 0 0.000 0.05 0.00 0.000 0.500 0.000 5.000 0.00 000: x 0.000 0.00 0.00 0.000 0.00 0.000 0.00 0.00 5005-05 0 0.00 0.00 0.00 0.000 5.00 0.000 0.00 0.00 000 a 5.00 5.00 0.00 0.000 0.00 0.000 5.00 5.00 050: a 5.000 0.00 0.00 5.000 0.00 0.000 5.50 5.00 00 x 0.000 0.00 0.00 0.000 0.00 0.000 0.00 0.00 000 x 000:0 0.000 0.00 0.00 0.05 0.00 0.00 5.00 0.00 000:0 x 0.000 0.00 0.00 0.000 0.00 5.000 0.000 0.00 0000 x 0.000 0.50 0.00 0.000 0.05 0.500 0.000 0.50 5005-05 a 0.00 0.05 0.00 0.000 0.00 0.000 0.00 0.50 <00 0 0.00 0.00 0.00 0.000 5.00 0.000 0.000 0.00 050: x 0.000 5.00 0.00 0.500 0.55 0.000 0.000 0.50 00 x 0.000 0.00 0.00 0.00 0.00 0.00 0.500 0.00 000 0 0:0 5.000 0.05 0.00 5.000 0.500 0.050 0.000 0.00 000:0 x 0.000 0.05 0.00 0.000 0.500 0.050 0.000 0.00 0:0 0 0.000 0.00 5.00 0.000 0.00 0.000 0.000 5.00 5005-05 0 0.000 0.00 0.00 5.000 0.00 0.000 0.000 0.00 000 x 0.000 5.50 0.00 5.500 0.00 0.000 0.00 0.00 0500 a 0.000 0.00 0.00 0.000 0.00 0.000 0.000 0.00 00 0 0.000 0.00 0.00 0.000 0.000 0.000 0.000 0.50 000 x 000: 0.000 0.00 0.00 0.000 0.00 0.000 0.00 0.00 000:0 x 0.000 0.50 0.00 0.000 0.05 0.500 0.000 0.00 0:0 0. 0.000 0.00 0.00 0.000 0.00 0.500 0.00 0.00 0000 x 0.000 0.50 0.00 0.000 0.50 5.000 0.50 0.00 <00 0 0.000 0.00 0.00 0.000 0.00 0.000 0.00 0.00 050: x 0.000 0.00 0.00 0.000 0.00 0.00 0.00 0.00 00 x 0,000 0.00 0.00 0.500 0.00 0.000 0.00 5.00 000 0 5005-05 0.000 0.00 0.00 0.50 0.05 0.000 0.50 0.00 000:0 0.000 0.00 0.00 0.00 0.00 0.00 0.000 0.00 0:0 0.000 0.00 0.00 0.000 0.05 0.000 0.000 0.00 0000 0.000 0.00 0.00 0.500 0.00 0.55 0.00 0.00 5005-05 5.850 com «o :x: ucme ucmHm AaavwvocuwucH 0000: 00 5800 .03 huwuaumz 003000 sumcmq mmmuw>< \mvom a \mwswomm 0 mo nuwcmq 0005:: .uoa hue 0:000 00 0000 000 0» 0000 A.u:oov .m @0009 133 mo. ~.H~ w.¢o~ h.- n.m m.~m Hue x «o. «.mHH q.amm o.on ~.m N.QMd mmoo x No. m.o~ o.~o~ o.oq m.c o.mm hmquumfi x Ho. n.0m o.eom m.om ~.e a.qu »a= \.u3 hug vmmm \oomom a .ua vummuooH vom\mumom ~ \.u: hue mom A.uco0v m manna 34 mmucmuwwmwa .me NH «« mmuaauowmfia .wfim Na « Hoe. mHo.CH “Houumv sum x gum «*amh.ma mHo. «*¢om.mo Hm¢.amo Ammazuoaoov mafiaflamm «ham.m «co. «gmqm.m ~¢q.mm acofiumUHHamM m> .m.z m> .m.= xmvcH umm>umm com mo numawa mwmum>< www.mqm mmo.¢ oma. ohm.¢NH Auouumv 8mm x 9mm {*Hm~.m mmw.mmma *«0¢H.nm mnm.onm «*oqa.oa Hmo.N «aqwm.m OHH.HMOH Ammnhuocoov mmwaaamm ««~Ho.m wnm.NMNH omH.H Nwm.¢ ¢«MNO.NH 5mm.H «*nmm.m wmm.omm macauwuwaamm M> .m.z m> .m.x m> .w.z m> .m.z C: ”.525“: in 3% Cu J3 ummmnooa Arv vom\mvwmm mo * Axv uamam\mvom «0 * nmm.quq Hm~.mw¢ mHo.mm oom.~m~ Auouumv 8mm x mom *«waw.oa mmo.o-~m ««wm~.< nqm.mwom «*qom.n Hom.wmm ««mmo.m~ me.qmqo Ammmauocmuv mmHHHamm «cqn.m cm~.omeH «*oao.q moo.nm- ««Hoa.q owm.qu ¥«wco.ma mma.mnm¢ mcowumowammm m> .m.z m> .m.z m> .m.z m> .m.z uamHm ucmHm mmvoaumuaH \mvmmm «0 * ucmam\.u3 %ua wom \mwaoumm mo § mo nuwcwq mmN Hmuoa mmq.ma~ moa.m~m qn~.ma mwm.¢ mmH Auouumv 3mm x mum *Nam.ofi Noo.o~aq ««~mo.m mqm.cqmq *«woq.hfi Hmo.am~ «gaqo.om mhm.~wa me Ammaxuoamov mmafiamm «Num.m mam.o~w ««n@n.n mom.cawo Hmm.~ ooo.am mmH.~ Hmn.oa m maoaumuwaamm m> .m.z m> .m.z m> .m.z m> .m.z mmvoz no * Hmuoa unmam\.u3 mun uamam mUHHSumz cu mhma wcaumsoam ou mama .m.v _ .mowumu mucmwuw> mo muamoamaawfim new mmumsvm name we zumassm .qnma a“ aways»; Hm can Hmucmuma .m x w wnu mo ooawaua> mo mamxama< .a manms 35 the early post-emergence stage. This is thought to have caused the delay in growth and development of the plant, resulting in delayed flower appearance. Twenty-two out of the 28 F hybrids had mean 1 values greater than that of the mid-parent value, and 8 hybrids had mean values greater than that of the late flowering parent. Six hybrids had mean values smaller than that of the early flowering parent. The hybrid that flowered the latest was the one derived from the cross BTS x 0685 (55.2) and its reciprocal, 0685 x BTS (57.5). The first hybrid to flower was the one derived from the cross SB x 72-7427 (28.9) and its reciprocal, 72-7427 x SB (28.4) (Table 3). Figure 2 shows the Wr/Vr graph of the 8 x 8, F1. The regression line is significantly different from zero but not significantly different from 1 (b = .88 .07). The line intersects above the point of origin (a = 6.13) indicating apparent partial to complete dominance as the JF§I7B value is 1.11. All the array points seem to fit the regression line very well, with parents 7 and 8 carrying a preponderance of the dominant genes since they are located near the point of origin and parents 2 and 5 containing a preponderance of recessive genes since they show both high variance and high covariance. The positions of the points representing BTS, 0674 and 0685 indicate that they contain neither an excess of 36 C 1 - BTS 0 680685 (012::53 l. 7=3TUI 60 F 0380674 . 8=JULES .4 = SEA 2 5 50 , O 5 3 72-7427 40 F .4 1 W' 30 + .9 6 20 + 08 7 b 8 .88 1.07 * 0 =613 : 12.39 ** 1O , o l I l 1 n _. 1O 20 3O 4O 5O 60 Vr * t for b = 0 = 13.54 and t for b = l = 1.78 *1: t for a = 0 = .48 Figure 2. Days to first flower (8 x 8, F1). 37 dominant genes nor an excess of recessive genes. Since the F value was -20.38 it is indicative of an excess of recessive alleles among the parents as a set. The herita- bility estimate was 32% (h2 = .32). Table 5 shows the mean values for the parents and the F2 families in the 1974 trial. Eight out of 10 F2 families gave mean values exceeding that of the mid- parent values and 4 out of the 8 F2 gave mean values exceeding those of the late flowering parents. Two F2 families had mean values smaller than that of the mid- parent values. The array means (Table 4) show that there are additive effects from the F1 to the F2 generations in all the families except those derived from the parent 0674. Figure 3 shows the Wr/Vr graph of the 5 x 5, F2. The regression of the Wr on Vr is neither significantly different from 1 nor zero (b = 0.6 i .70). The line intersects the Wr axis above the origin (a = 7.17 i 12.86) and significantly different from 0, indicating the genes controlling days to first flower are in the partially dominant range, on the average. The evidence on the graph indicates genic interaction. In Figure l, the position of parent 3 (0674) was near the point of origin indicating that it contained the dominant genes, 38 m.oq o.om m.moa ¢.wm m.om ~.oo w.¢m ame .m manna 39 on. m.sq m.ooH m.- w.m o.mn m.moH «mm x Nm. H.me H.HwH q.o~ m.e H.mo m.mHH «moo x mm. «.mm N.OHH o.oq n.m m.eo s.w~H mm x mm. ~.~q ~.mma N.n~ o.m n.~o m.~oe mam x emseusm em. ~.es m.omH H.om N.s H.~o m.OHH emseume x mm. H.Hm w.~HN o.¢H s.¢ m.H¢ s.~m «use x mm. H.~o m.o- q.e~ a.m «.me «.sm mm x mm. o.mo o.smm 5.0H o.m N.H~ N.wm men x umm. \.03 sun emmm \memmm e emmmnooa \memmm e \.uz sun see no euwamg A.u:00v .m mange 40 .1 =91: O 2:33 30 . . 3=0674 Cb4==SEA O 53 72-7427 .4 3 20 * ‘. VVr ‘15 4.1 .2 10b b=-60:.70* a :3 117112.86 ** o - 1:. 10 2O 30 Vr II C * t for b *1: t for a II C Figure 3. .86 and t for b = l = .58 = .56 Days to first flower (5 x 5, F2). 41 however, in Figure 3, this point shifted to the far right. The data here appears to be inconsistent with previous interpretation, due primarily to strain 0674 having switched its position drastically from near the origin in the F to the far right in the F2. Parent 4 1 (SEA) shows its consistency regarding this trait in that it continues to behave as though it contained a prepon- derance of recessive genes controlling days to first flower. BTS, SB and strain 72-7427 also were positioned consistently on the graph in both generations. Even though the heritability estimate was 15% for the F2, the h2 values in the 5 x s, F was 29% and 32% for the l 8 x 8, Fl; the evidence from the "D" values in the three sets of data suggests that days to first flower is mod— erately heritable and can be transferred to the progenies. "F" has a negative value of -27.32 indicating that most of the parents contain recessive allels. The estimates of additive variance (D) in both generations were low (D = 15.22 in 5 x 5, Fl' D = 21.70 in 5 x 5, F2). How— ever, the array means (Table 6) show that there are additive effects in all the families except those derived from strain 0674. Therefore, on the basis of the array means, it is not at all impossible to select parents for making crosses to obtain progenies with either late or early flowering habit. 42 mocmumMMHa .me NH «« mmm.~H oqo.mm Nu Auouumv 8mm x mom ««o~¢.c nmq.mm *«oao.mw www.maoa «N Amoezuocmov mmHHHamm new. man.m wmm.m mwm.omN m msowuwowadwm M> .m.z m> .m.z xooaH umopumm pom mo nuwcmq mmq.mn mmm.~ woo. «wo.Hc Nu Auouumv 3mm x mom «*Hmm.m qoo.omm ««~w¢.me me.voq «tmmm.- mom.H «#omm.ma mq~.mom em Amoazuoaouv moaawawm «humo.ma mmm.oHHH mmo. nmq.H «« .m.z m> .m.z m> .m.z m> .m.z “xv uanm\.u3 mun comm ANV .uz omomnooa va oom\mvmwm no a .Axv nomam\moom mo e qu.eeaa mHo.mm o-.m mmo.Ho~ Nu Auouumv 8mm x mom «teen.oa mmH.monH ««Hom.m mam.on ««~nm.~a omw.aaa «snow.~H coq.¢nn~ «N Ammahuosoov moaaafimm *«omn.ma enn.~mowa «tmqa.ma www.mmwa «xwmn.n mm¢.ob «*qu.m~ eme.mamn m mcofiumuaaaom m> .m.z m> .m.z m> .m.: m> .m.z uomam usmam motocuoucH \mpomm «0 a ucmHm\.u3 hum com \moaoomm mo * mo Suwamq ooo.mn wNm.o~H «Hw.o mmm.~H Nu Auouumv sum x mom aa~m~.me nem.eom aamee.m mee.ema “amme.me wmm.eoH «aoea.e eas.mm ea Amoesuoaaov amaeaaae «xomo.w mqo.m~q ttaam.m~ oma.qmoq *«Hem.q wom.mm mow. man.m m macaumofiaama m> .m.z m> .m.z m> .m.z m> .m.z mowoz mo e Hmuoa .u3 mun ucmam huwusumz on axon wafiuoaoam ou mhmo .w.v .moaumu mo .qmofi ca mGOHumumcmw cwwum> mo cosmowmwcwfim vow moumavm sums mo %Hmaasm mm was Hmucoumu .m x m ocu mo oucwaum> mo mam>au¢< .0 manna 43 Duration of Flowering 5 x 5, Fl The analysis of variance showed that there were no significant differences among replications but there were significant differences among the genotypes in dur- ation of flowering (Table 1). Strain 0674 was the variety with the longest mean duration of flowering of 28 days. Strain 72-7427 had the second longest mean value of the trait of 24.7 days. SB and SEA had about the same mean value of the trait of 22.8 and 22.7 respectively. BTS was the variety with the shortest mean duration of flowering of 20.4 days. Six out of the 10 F1 hybrids gave a mean duration of flowering longer than that of the parents. Those were the crosses involving BTS x 0674, BTS x SEA, SB x 0674, SB x SEA, SB x 72-7427 and SEA x 72-7427. It is noticed from the six crosses mentioned that the varieties BTS and SB produced an F1 with heterotic effect only when used as pistillate parents whereas strain 72-7427 and strain 0674 gave hybrids with heterosis when used as the pollinate parent. SEA, how- ever, was not consistent regarding this trait. Figure 4 shows the Wr and Vr graph for the 5 x 5, F1. The regression line is not significant (b = .17 i .37). The simple additive effect model cannot be applied to this set of data. The computed "D" value was 10.40 and the average degree of dominance was 1.58; since the 44 an. noun .0.“ .Aam .m x mv msflmosoam mo soflumuso .q musmwm so muéuwun .2 p “Vol Dun uOu a ._> on me o. m m a. .0 «345004.. a hm.“ hr. u n :2. «sum 0 N. (mnuc C goons 0 menu 0 when r O :5 Op 45 array points are so widely scattered in Figure 4, it is concluded that the data may not be reliable. The heritability estimate was 15%. As to the positions of the array points, parents 1 and 5 appear to show a certain level of dominant genes, parent 2 appears to contain the recessive genes. Parents 3 and 4 seem to be the outliers, particularly parent 4 (0674). It was discussed earlier that strain 0674 was very much affected by ozone injury. Strain 0674 resumed its growth and development later in the season when cli— matic conditions were favorable again. It continued to produce flowers. However, the season was not long enough for its second phase of development; most of the flowers developed into small pods which were not fully filled. This trait was not measured in 1974. Days to Maturity The analysis of variances of days to maturity showed that there were no significant differences among replicates but highly significant differences among the genotypes (Table 2). Inspecting the mean over replicates in Table 1, it appears that SEA was the earliest and strain 0674 was the latest maturing variety with mean values of 74.8 and 92.6 days respectively. Strain 72-7427 and varieties BTS and SB had mean numbers of days to maturity of 85, 89.4 and 89.5 respectively. 46 .It was surprising to see that SEA and strain 0674 which tuive several morphological characters in common did not behave similarly regarding this trait. However, the differences between the two means was only about 8 days. It: was observed that both 0674 and SEA were affected by ozone injury during the pod-filling stage. The leaves were senesced prematurely and abscised. Therefore, ruarmal photosynthetic activities were disrupted and the pods were not fully filled. Strain 0674 appeared to be rmore affected by ozone injury. After this period, due to favorable climatic conditions, strain 0674 resumed its growth and development but not SEA. The flowers that strain 0674 produced at this later stage delayed its <1uration of flowering and maturity. The other pair of parents which have several ntrolled largely by dominant genes and early maturing is controlled largely by recessive genes. It can be observed that without the presence of the 3 new parental lines, SB has retained its position on the dominant side. The additive genetic variance (D) has the value of 42.38. The heritability estimate is 14% for days to maturity which is relatively low. The "F" value of 43.42 indi- cates an excess of dominant genes among this set of 53 O1=BTS OTSTUI 02:88 .BsJULES C3=0674 .4 = SEA 60r 05:72-7427 06:0685 .16 * 2.16 ** 60 * ‘tfarb ll 0 II 7.28, t for b = 1 II I H O *9: 'tfara II C II 2.03 Figure 7. Days to maturity (5 x 5, F2). 54 parents for this trait. The F2 populations derived from BTS, SB and 72-7427 seemed to be reasonably uniform in maturity. However, those derived from SEA and 0674 showed several days difference in maturity, particularly those resulting from 0674 and SEA and its reciprocal. Ozone injury is thought to be the cause, since both 0674 and SEA were very sensitive to this environmental factor. Planet Dry Weight 5 x5,Fl The analysis of variance shows that there were no significant differences among the replicates but there were highly significant differences among the genotypes. Different plant types certainly contributed to the dif- ference in this trait. There was no significant dif- ference between the F1 and its reciprocal. Among the 5 parental varieties, BTS had the highest mean plant weight of 100.9 gm and 0674 had the smallest mean plant Weight of 51.0 gm (Table 1). SB, SEA and 72-7427 had mean values of 68.9, 55.9 and 78.3 gm, respectively. Eight out of 10 hybrids had mean values exceeding that Of the highest parent in the cross. Two hybrids had mean values lower than that of the mid-parent values. BTS x 0674 and its reciprocal produced hybrids with highest mean plant weights of 202.6 gm and 195.4 gm, respectively. The hybrid with the lowest mean value derived from the cross 0674 x SEA (45.2) . 55 It can be seen in Table 7 that BTS has the highest array mean in both F1 and F2. All the array means decreased in the F2 generations. The heritability esti- mate was -.16 which has to be considered as zero. The large negative estimate for "D" of -3385.36, indicates that there was a very high error variance. Therefore, further interpretation will not be reliable. It is assumed that this set of data does not fit the model. However, the Wr/Vr graph presented in Figure 8 indicates tfliat the regression line is neither significantly dif- ferent from b = 0 nor b = l (b = .34 r .26). The line intersects below the point of origin but not significantly different from 0 (a = -452.10 1r 561.83) . The intermediate Slope value is indicative of genic interaction being involved in the expression of total plant weight. Genic interaction, as deduced from the Wr/Vr graph, merely obscures the manifestation of genic additivity or domi- nElnoe. It does not, in itself, exclude them. More eS-Eplanation will be given in the section of "Discussion." The position of the array points indicates that 8train 72-7427 and variety SB contain a preponderance of recessive genes. Variety SEA seems to contain a balanced Proportion of dominant and recessive genes. BTS, which is the only indeterminate type among the 5 parental lines, behaves as an outlier. It has the highest mean Plant weight, as stated earlier. It is not surprising 56 .xam .m x me so as sesame who ocean .m musmam om. n o u a How e #* Hm.~ u H u a How u .om.a u o u a you as o . I . aamo_bm + crane: u a . .> no ooov coon ooou .oo. o .23 e O h>3 (mm .1- Q C . COOP no @500 u n 0 mm P. N . gOOVP 33.1: n m 0 men n F o 57 that its position on the graph will be far away from the others. The contrasting phenotypes and interacting loci would cause this difference. 8 x 8, Fl The rankings of plant dry weight in grams of the parental varieties in the 1974 trial from the highest to lowest mean values were as follows: strain 0685 = 118.1, Jules = 101.8, Tui = 99.6, BTS = 95.7, SB = 89.5, 72-7427 = 77.9, SEA = 64.6 and strain 0674 61.8 (Table 3). Twenty- three out of the 28 F1 hybrids had mean plant dry weights higher than that of either parent. The means of 4 hybrids were in between the two parents and the mean of one hybrid (Tui x Jules) was lower than that of the lower parent in the cross (Table 3). The cross involving BTS x 0685 and its reciprocal gave hybrids with the highest mean plant dry weight of 241.6 and 222.6 gm, respectively (Table 5). The hybrids with the lowest mean plant dry weight derived from the cross 0674 x SEA and its reciprocal (62.9 and 73.6 gm respectively). The Wr/Vr graph (Figure 9) shows that the regres- sion line is significantly different fromb = l and it is not significantly different from b = 0 (b = .11 i .13). The simple additive gene system cannot apply to this case. The estimate of "D" also has a negative value of -325.94 as with the 1973 data. This is probably due to the high 58 . H o How u A m .m x we Eb cw unmwo3 mum uqum .m ousmam 6H.H u o n e. .> eH.H n a u h hoe 6 .me. u e n n how he com. cow. own _h a com. mmdafi .3». name NNVNINN h. mm. (mm #500 wk“ .3 v. .ooop n. N. 59 error variance associated with this trait (E = 875.10). Therefore the assumption of no gene interaction is not valid in this case. It is observed that the arrays which show interaction are BTS and strain 0685. If these two array points are omitted the remaining arrays may suggest a different slope, i.e. a slope of b = .7 is possible and a different interpretation of the mode of gene action is expected. (Strictly speaking, if one or another par- ental array is to be excluded from the Wr/Vr graph, a new regression line should be calculated omitting all data involving the excluded parents. In the present instances, an approximate line has been fitted by eye to the original array points, without recalculating the array variances and covariances.) 5 x 5, F2 Table 5 shows that 6 out of 10 F2 populations (excluding their reciprocals) had mean plant weights greater than that of their heavier parent, 3 populations had their means in between that of the two parents and 1 family gave a mean value smaller than that of the lighter parent. These mean values were all smaller than those of the F1 generations due to a lower level of hetero- ZYgotes (Table 4). The Wr/Vr graph in Figure 10 shows that the regression line is significantly different from b = 0 60 .Amm .m x my Em aw psmfim3 Mus unmam .oH musmflm Mhohm " O H .0 HOW u. Nfl.Nhl ts NN.I u H u Q How u .mm.oa u o u Q now us .> can con ecu co. «6 . r. on; u 3.2.: u a e 3. w «o; u n . on. :s Lecu «see u o o v 31...: u m 0 am I a o no (umuoo meunpo .oou 61 but not significantly different from b = l (b — 1.02 t .10). The line intersects the Wr axis significantly below the point of origin (a = -72.3 i .86) indicating over-dominance. This agrees with the interpretation from the 5117:) value of 4.05. The fact that BTS, SB and strain 72-7427 are located near the point of origin indi- cates that these parents contain mostly dominant genes controlling this trait and strain 72-7427 has most of the dominant genes among the 3 parents (Figure 10). Strain 0674 and SEA contain mostly recessive genes as indicated by their positions on the top right of the graph. There is not much difference, however, between the array means of these 2 groups of parents. It may be assumed that plant dry weight is influenced both by dominant and recessive genes. The heritability estimate is essentially zero (h2 = .04). However, the mean values in the 8 x 8, F1 indicated that the heaviest hybrids derived from the most vigorous plants and the lightest hybrids from smallest parents (Table 3); it is, therefore, believed that this trait is heritable and can be transferred to selected families in the next generation. Number of Main Stem Branches 5 x 5, F l The number of branches on the main stem were counted at maturity and recorded on the individual plant basis. Table 1 indicates that strain 0674 bore the 62 highest number of main stem branches with a mean value of 10.08, SEA, 72-7427, BTS and SB had mean values of 9.8, 9.3, 7.9 and 7.5, respectively. Six out of 10 F1 hybrids gave mean values greater than the mean value of the high parent. Two hybrids had mean values greater than that of the mid-parent value and 2 hybrids had mean values smaller than that of the low parent. It is inter- esting to observe that when both strain 0674 and SEA were crossed with other varieties all the hybrids showed heterosis except when these two parents were crossed to each other. It appears that strain 0674 and SEA are similar genetically. The regression of Wr on Vr in Figure 11 shows that it was not significant from zero (b = .29 i .30). Genic interaction is indicated. The assumption of no gene interaction is not valid. However, examining the array points indicates that SEA contains the highest number of dominant genes. BTS and strain 72-7427 appear to contain a certain portion of dominant genes. SB contains the most recessive genes regarding this trait. Parent 3 (0674) is the outlier in this case and causes the slope of the regression to deviate from unity. As it was mentioned earlier, strain 0674 resumed its growth after being affected by ozone injury. This genotype and environmental interaction in the later stage of develop- Inent probably increased error variation, causing the 63 H .Aam .m x my monosmnn Eoum same mo Honfisz .HH ousmwm i“ u a hoe u .mm. u o n n how u ¥ 0. 06 o.» O.N o; c. ‘ I d d o Oh. $0.." a «3% P. a “.0 ¥ on. a“. "n m. . 0.9 . m... hflwhuflh "m. a .35 a... . 6.“ C @500 no. mmuuo .m.« mhm N P. 64 negative value of "D." This is the trait that would be most strongly affected by density differences due to differential loss of plants in plots. Number of Main Stem Nodes 5 x 5, Fl The analysis of variance shows that there were no significant differences either among the replicates or the genotypes (Table 2). BTS had the highest mean number of main stem nodes with the value of 7.3, strain 0674, SEA, SB and strain 72-7427 had mean values of 5.4, 5.0, 4.2 and 4.1, reSpectively. No significant differences between the F1 and its reciprocal could be detected. Four hybrids had mean values greater than the mid-parent value and 3 had mean values exceeding that of the parent with the higher number of main stem nodes. Three hybrids had mean values smaller than the mid-parent values. Figure 12 shows the Wr/Vr graph of the number of main item nodes, 5 x 5, F1' The regression line was not significantly different from unity but significantly dif- ferent from zero (b = .86 i .24). The position of the array points representing SB, strain 0674, SBA and strain 72-7427 shows that they all possess recessive genes con- trolling this trait, with SEA being the extreme in this reSpect. BTS appears to contain more of the dominant genes. The evidence from the wr/Vr graph seems to 65 .1 " BTS 1.5" .23 SB .3 3 0674 .4 04 "SEA 0 5 3 72-7427 5 2 1.0» o 3 “h o 5 _ b = 66 t. .24 * .1 a =.2o: .24 ** o - l 0.5 1.0 1. Vr * tforb=0=3.62,tforb=l=.61 ** t for a = 0 = .82 Figure 12. Number of main stem nodes (5 x 5, F1). 66 indicate that the indeterminate type (BTS) contained the dominant genes and the determinate types (SB, strain 0674, SEA and strain 72-7427) contained most of the recessive genes controlling number of main stem nodes. Nevertheless, the estimate of h2 and D had values of -.23 and -17.24 respectively indicated that there was a large error variance (E = 19.61) in this set of data. There- fore, no interpretation can be made with confidence with regard to dominance and additive effects of the genes controlling this trait. Because of insignificant differences among the genotypes regarding this trait it was decided to determine the total number of nodes of the whole plant in the 5 x 5, F and 8 x 8, F populations. 2 1 Total Number of Nodes 8 x 8, Fl Total number of nodes consisted of all the nodes counted on every stem. This should give a better picture of the phenotype of the parents and that of the segregat- ing populations. Analyses of variance show there were no significant differences among replicates but highly sig— nificant differences among genotypes (Table 4). The branchy vigorous determinate type of strain 0685 gave the highest mean value of number of nodes (74.6) and the determinate, strain 72-7427 gave the lowest mean value 67 of 24.4. Jules, BTS, Tui, strain 0674, SBA and SB gave mean values of 72.2, 57.5, 56.4, 41.8, 39.9 and 30.9 nodes, respectively. Heterosis was strongly expressed in total number of nodes. Eighteen had mean values exceeding the mean values of the parent with large mean number of nodes and 6 had mean values greater than the mid-parent values. However, 4 hybrids had the mean values smaller than the parent with smaller mean (Table 3). The hybrid derived from BTS x 0685 gave the highest mean value of 196.5 (Table 3). The regression of Wr on Vr (Figure 13) shows that the line was significantly different from unity but not significantly different from zero (b = .16 i .07). The regression line intersects the Wr axis above the origin (a = 130.4 1 19.88) and is significantly dif- ferent from 0, indicating genic interaction. The position of the array points representing BTS and strain 0685 obviously caused the regression line to deviate significantly from unity. If we omit BTS and strain 0685 and draw a new regression line (dotted line) the remaining parental arrays would fit this line fairly well. It is apparent, then, that Tui contains a certain degree of dominant genes and strain 0674 contains the recessive genes. The other parents, SB, SEA, strain 72-7427 and Jules appear to contain an intermediate proportion of dominant genes. 68 .Aam .m x we mowoc mo Hogans Hmuoe .mH musmflm mm.e u o u a mom 0 ¥« mm.NH H H M Q HON # .N¢.N H o H D HON pr .> e.o \ com. come 666. com -\ o t. no.3 u 3.2: \ * uofiwoe.u.. \\ .aaa mwaafi n a 0 (mm 8 V O . 00°F .3». n b O QBOO n m C 0006 u D 0 mm .1. N C NNVNINN n m 0 who u F O 69 It is not clear why parents 1 (BTS) and 6 (0685) were positioned on the far right of the graph. They both had very high variances but low covariances. They looked as if they contain a preponderance of recessive genes which agreed with the estimate of F (= -720.46). Genotype-environmental interaction involving these two parents is probably the cause of this deviation. The heritability estimate was practically zero (h2 = .03). 5 x 5, F2 The total number of nodes of the whole plant was counted among the 5 x 5, F2 populations instead of count- ing the number of main stem nodes as in the 5 x 5, F1. It was found in the analysis of variances (Table 6) that there were significant differences among replicates and among genotypes. The mean values of the parents agreed very well in rank with the mean values for number of main-stem nodes in 1973. BTS with a mean of 56.5, strain 72-7427 with a mean of 26.7 were the highest and lowest, respectively (Table 5). SEA, strain 0674 and SB had mean values of 41.6, 40.0 and 33.5 nodes, respectively. Six out of 10 F2 populations had mean values exceeding that of the high parent. Two F2 populations had mean values greater than the mid-parent values but smaller than the high parent. Two F2 populations had mean values lower than that of the mid-parent value. 70 Figure 14 shows the Wr/Vr graph of the 5 x 5, F The regression of Wr on Vr is not significant (b = 2. .37 i .30) as was the case for 8 x 8, F1. Hence, there was no real relationship between Wr and Vr. The line intersects the Wr axis above the origin, (a = 43.33 t .86) and significantly different from 0, indicating genic interaction. The estimate of F was -4l7.18 indicating that there was an excess of recessive genes for this trait in this sample of lines. According to the position of the array means, parent 4 (SEA) contained a preponderance of recessive genes. None of the varie- ties seemed to contain a preponderance of dominant genes. Parent 1 (BTS) which differed phenotypically and geno- typically from the others was the outlier causing the regression to deviate from unity. Number of nodes is a complex trait. It is well understood that this trait effects other sequential traits, i.e. number of leaves, number of racemes, number of pods, etc. It is highly influenced by stand and density. BTS is an indeterminate type of bean. If occasional plants within plots are missing during vege- tative development, BTS could extend its growth in other directions, hence, producing longer stems, higher number of nodes, etc. Being of bush type, other parents did Inot have much extension of growth and development to (effect number of nodes. Therefore, those parents SB, 71 .> . Ame .m on my mmwoc mo Hon—sac Hmuoa .va mnsmfim ovw oou omp cow 0% o V0.0m H O n 0 50* a Id up an .0. unmuélonn .2. .OO Po QQ.HNO.MV "I u; on. .+. S. u a no .00.. no «0 VsoounO v0 hNQh.Nhu m 0 (mm! C. amid. iomw who fir. 72 strain 0674, SEA, strain 72-7427 position well within the parabola except strain 0674 which is slightly below the unit regression line. The heritability estimate in the F2 population is low (h2 = .04). Length of Internode 8 x 8, F1 The length of internode was not measured in the 1973 trial therefore no data for 5 x 5, Fl were available. In the 1974 trial it was decided to measure the length of internode. This was to provide a morphological picture of the individual line as well as for the segregating generations regarding this trait. Three of the longest internodes in each plant were measured and averaged to be the length of internode of that plant. Table 3 shows the mean length of internode of the parental and F hybrids derived from the 8 x 8 crosses. 1 It appears that the determinate type of dry bean has the longest internode (0685 = 195.6 m.m.) and the indetermi- nate type has the shortest (BTS = 89.0 m.m. Table 3). Of the 28 F hybrids, 17 had means exceeding mid- 1 parent values and 12 had mean values exceeding that of the long internode parent. Eleven F1 hybrids had mean values smaller than that of mid-parent values. The hybrid derived from the cross 0685 x SB gave a mean value of 73 248.4 which was the highest. Tui x Jules gave the hybrid with the lowest mean value of 79.2. Tui x BTS also gave a hybrid with short internode (94.2). The Wr/Vr graph presented in Figure 15 shows that the regression line is significantly different from b = 0 and b = l (b = .51 i .04) indicating genic interaction plays a major role in determining length of internode. Balancing the evidence on the graph, particu- larly the intersection of the regression line on the Wr axis, it is believed that genic interaction is involved and the calculated value of / Hl/D = 1.2081 should not be relied upon as indicative of the prevailing mode of gene action. The positions of the array points show that Tui, Jules and BTS contain a preponderance of dominant genes, strains 0674, 0685 and the variety SB contain a prepon- derance of recessive genes. SEA and strain 72-7427 appear to contain a balanced proportion of dominant and recessive genes. Since the "F" value is 1386.33, indicating an excess of dominant genes among the parental lines, SEA and strain 72-7427 may contain a higher pro- portion of dominant than recessive genes. The array means of strain 0685 and SB were the highest (183.9 and 174.8, Table 8) and the positions of the two parents are on the recessive side; it is therefore reasonable to believe that these two parents will produce offspring 74 .Aam .m x we ES ca moocuoucfl mo season .mH ousmam .> ooou ooo. ooo. com -o «oo a cum to. . e.o Sam. at... to. 2 v0.3 noun .0. a . com o. .2. 8833.8? 6 e a 3.“ end a no o o .ooo. o0 no nmaoauoo <3 .30 Shun. whoonn. oooouoo oouuo loom. :5 32333.... who... 75 with long internodes if they are crossed and that long internodes are controlled by recessive genes. BTS, Tui and Jules are all indeterminate type. According to their array positions they all contain dominant genes. There- fore, it is also reasonable to assume that short inter- nodes are controlled by dominant genes. 5 x 5, F2 Among the 5 original parental lines, strain 72- 7427 had the longest mean internode length of 155.7, SB, strain 0674, SEA and BTS had mean values of 150.1, 129.4, 113.4 and 88.6 respectively (Table 5). The rank— ing of the 5 parental lines here are not in the same order as found in 8 x 8, F (Table 3). They were, in fact, 1 the same group of parents, grown in the same field. The differences in ranking were due to the number of means taken to compare with their corresponding F1 (Table 3) or F2 (Table 5). In Table 3, only 2 parental means were averaged whereas in Table 5, because there were 6 F2 plots, 6 parental means were randomly taken from the overall 14 means to average and compare with the F2 means. The mean value of strain 72-7427 in Table 5 is larger than its corresponding mean value in Table 3 simply because the 6 parental means which were randomly taken happened to have large values and the 2 means taken to be averaged as the parental mean in Table 3 happened 76 to have small values. Since the parental means in Table 5 were averaged from a larger number of values, therefore, they are more reliable. Six out of 10 F2 families examined had mean values exceeding that of the mid-parent values and 3 out of 6 F2 had mean values larger than that of the parent with long internode. Four Fz's had mean values smaller than that of the mid-parent values. Three out of 4 F2's mentioned derived from crosses involving BTS as the maternal parent and 1 derived from the cross SB x 72-7427. The F2 family with the longest internode was the one resulting from the cross 0674 x 72-7427 (166.9, Table 5) and the F2 family with the shortest internode was the one derived from BTS x 0674 (105.4, Table 5). The Wr/Vr graph presented in Figure 16 shows that the regression line is significantly different from b = 0 but not from b = l (b = .68 i .24). The intercept on the Wr axis is above the origin but not significantly dif- ferent from 0 (a = 132.25 1 121.32) indicating complete dominance. The / Hl/D value of 1.98 which indicates over dominance does not agree with the interception of the regression line. It is obvious from the graph that the interception is above the origin and the array points fit the line very well. Therefore, the calculated value of Hl/D is not believed to be reliable and the 77 .ANm .m x me SE CH opocuoucfl mo awocwq .ma ousmflm 5) 00v OON 000 040 1 1 O OO.PIO"Q 50* a . . n u so a «néuwuntoe «.hhu O n u w .00“ «0 No.3.“ mueopuo .OOQ VN.“ Do.un_ :5 m. «0 .OO@ choc fl 0 C 0. hNQh-Nh n m C mm H N . (mm-"Q. mhfllu P. 005 78 interpretation probably should be done strictly from the evidence on the graph, which suggests partial to complete dominance. The array points indicate that BTS contains a preponderance of dominant genes. Strain 0674 contains a preponderance of recessive genes. Strain 72-7427 con- tains a higher proportion of dominant and recessive genes. The "F" value of -743.48 indicates that there is an excess of recessive genes among the parental lines. Since the positions of more points are on the recessive side, the "F" value confirms the picture on the graph. BTS (parent 1), which is the only indeterminate type among the 5 parents, possesses dominant genes. Its position in the 8 x 8, F1 is also on the dominant side together with other two indeterminate lines (Tui and Jules). This consistency shows that short internodes are controlled by dominant genes. The heritability estimate is 14% which is low. This indicates that length of internode is not highly heritable in the F2. It would be difficult to select efficiently in the F2. It also suggests that in F3 or advanced generations selection for this trait should be on a family basis and not on an individual plant basis. 79 Number of Racemes Per Plant 8 x 8, Fl Number of racemes was not counted in the 1973 trial. In 1974 this trait was included and it was deter- mined from both the F1 and the F2 populations. Among the 8 parental lines, it appears in Table 3 that strain 0685 had the highest mean number of racemes (30.6) and strain 72—7427 had the lowest mean of 11.5. Twenty-four out of the 28 F '3 had mean values greater 1 than that of mid-parent values and 21 out of that 24 Fl's had means greater than the mean of the high parent. Four crosses resulted in Fl's which had mean values lower than that of their mid-parent values. They were: SB x 72-7427 (14.4), 0674 x SEA (19.8), SEA x 0685 (23.5) and Tui x Jules (22.5). The cross involving SB x 72-7427 and its reciprocal gave hybrids with the lowest mean number of racemes of 14.4 and 15.0 respectively. The hybrid derived from the cross BTS x 0685 gave the highest mean value of 54.9. It is also shown in Table 8 that the array means of strain 0685 and BTS were among the highest (34.6 and 34.1). The situation presented by the Wr/Vr graph (Figure 17) is confusing. The regression is shown to be significantly different from b = l but not significantly different from b = 0. The simple additive gene system cannot be applied to this case. Furthermore, the computed 80 ‘.1==BT55 695:172-7427 02:88 06=0635 C3=0674 .7=TU' .F4=§3EA1 ‘88==Jl’LEEB 8()' / ' Wr 40 2(D’ o A ’//r so 100 Vr _20 b=.34.t.17 1. a 23.881 10.36 ** -4C) * . t for b II 0 II 2.04 and t for b = l = 4.01 * II C II *t for a 0.37 Figure 17. Number of racemes per plant (8 x 8, F1). 81 "D“ has a negative value of -2.86. Gene interaction is believed to be involved. Therefore this set of data does not fit the "no gene interaction" assumption for the diallel model. The positions of the array points look disturbing particularly those representing BTS, strain 0685 and SEA. If a new regression line is drawn (dotted line) as shown by omitting BTS, strain 0685 and SEA, the line would fit other array points and would be signifi- cantly different from b = 0 but not from b = 1. From this new position of the regression line, SB, strain 0674 and strain 72-7427 may be said to contain recessive genes and Tui and Jules contain a moderate level of domi- nant genes. However, the interpretation from this set cannot be considered reliable since "D" has a negative value. 5 x 5, F2 Nine out of 10 F2 populations examined had mean values exceeding that of the mid-parent values (Table 5). The only population that had a mean value lower than that of the mid-parent value was the one derived from SB x 72-7427 (14.1). The cross BTS x SEA appeared to give an F2 population with the highest mean number of racemes. The F2 derived from the cross SB x 72-7427 had the lowest mean number of racemes. The array means (Table 8) also show that the lowest values belong to SB and strain 72-7427. 82 Table 7. The array means of different traits for the F and F2 (5 x 5) taken from Tables 1 and 3. 1 BTS SB 0674 SEA 72-7427 Days to First Flower F1 36.0 31.4 39.7 34.0 28.3 F2 42.4 35.3 39.9 36.7 33.7 Days to Maturity F1 95.8 92.4 92.1 84.3 84.6 F2 95.4 92.6 92.6 87.5 91 2 Plant Dry Wt. F1 128.9 109.6 107.3 103.9 86.7 F2 93.7 88.9 88.0 82.7 84.7 # Main Stem Branches F1 8.4 9.7 10.1 9.6 9.3 # Main Stem Nodes Fl 7.3 5.2 5.9 5.9 4.6 Pods Dry Wt./Plant F1 102.2 83.7 77.3 81.9 68.2 F2 69.2 65.5 61.8 61.5 63.4 # of Seeds/Plant F1 406.7 230.5 318.8 288.4 144.9 F2 255.1 165.4 239.8 229.8 144.7 # of Pods/Plant (X) F1 74.4 52.1 69.7 60.3 32.1 F2 52.3 42.4 53.2 50.8 34.2 # of Seeds/Pod (Y) F1 5 l 4.2 4.5 4.5 4.3 F2 5 0 3.9 4.3 4.3 4.1 loo-Seed Wt. (2) F1 21.9 32.3 18.0 21.5 47.9 F2 21.0 33.8 20.0 21.6 36.4 Seed Dry Wt./Plant F1 76,9 66.0 58.6 63.1 50.9 F2 53.7 51.1 46.9 49.1 47.9 Harvest Index F1 .60 .60 .55 .60 .59 F .58 .59 .54 .57 .57 2 83 N 6.66 6.66 6.66 6.66 6.66 66 6.666 6.66 6.666 6.66 6.66 6.66 6.66 6.666 6 6:66m\.uz mun 606 6.66 6.66 6.66 6.66 6.66 W6 6.66 6.66 6.66 6.66 6.66 6.66 6.66 6.66 6 666 66666\6666 6 6.66 6.66 6.66 6.66 6.66 6.66 6.66 6.66 66 ucmam\moaoomm 6 6.666 6.666 6.666 6.666 6.666 6.666 6.666 6.666 66 moosuoucH mo :uwcog 6.66 6.66 6.66 6.66 6.66 M6 6.66 6.66 6.66 6.66 6.66 6.66 6.66 6.66 6 60602 mo 6 66608 6.66 6.66 6.66 6.66 6.66 m6 6.666 6.666 6.666 6.666 6.666 6.666 6.666 6.666 6 .uz mun nomam 6.66 6.66 6.66 6.66 6.66 66 6.66 6.666 6.666 6.66 6.66 6.66 6.66 6.66 66 zuausumz ou meme 6.66 6.66 6.66 6.66 6.66 M6 6.66 6.66 6.66 6.66 6.66 6.66 6.66 6.66 6 603066 666 on 6669 66666 666 6666 6666166 666 6666 66 666 .m can H moanwb aoum coxmu Aw x we mm won am 056 now 666muu econommat mo memos 60666 one .m oflemH 84 66. 66. 66. 66. 66. M6 66. 66. 66. 66. 66. 66. 66. 66. 6 wovcH umo>umm 6.66 6.66 6.66 6.66 6.66 m6 6.66 6.66 6.66 6.66 6.66 6.66 6.66 6.66 6 666666.63 666 6666 6.66 6.66 6.66 6.66 6.66 “6 6.66 6.66 6.66 6.66 6.66 6.66 6.66 6.66 6 666 .63 66661666 6.6 6.6 6.6 6.6 6.6 M6 6.6 6.6 6.6 6.6 6.6 6.6 6.6 6.6 6 666 666666666 6 6.666 6.666 6.666 6.666 6.666 m6 6.666 6.666 6.666 6.666 6.666 6.666 6.666 6.666 6 . UCNHW\WUO@W «x 6.666 6.666 6.666 6.666 6.66 6.66 6.666 6.666 66 666 66 666666 66666 666 6666 6666166 666 6666 66 666 6.66666 .6 66666 85 Figure 18 shows that the regression of Wr on Vr is significantly different from b = 0 but not from b = 1 (b = 1.11 i .23). The intercept is below the origin but not significantly different from O (a = -2.76 i 4.496) indicating that the dominance is complete. The / Hl/D value, however, was 2.1733 which indicated over dominance. The evidence from the graph and /F§;7B is suggestive of complete to slight over—dominance. The position of the array points show that SB contains a preponderance of recessive genes. SEA and strain 72-7427 contain a lesser proportion of recessive genes. Strain 0674 and BTS appear to contain a balance proportion of dominant and recessive genes. The position of the points agree well with the "F“ value of -39.6466 which indicates that there is an excess of recessive genes among the parental lines. The heritability estimate is 11% which is not very high for this trait. Since additivity of the gene system was not indicated, selection for this trait may be difficult in the present sample of lines and their intercrosses; the gene action picture, however, does not imply that great difficulty would necessarily be encountered in a larger sample of parental lines. 86 ‘. 1"BTEB O 2358 30’ . 3=0674 C 4=SEA C 5:72—7427 20" b 61.11:.23 * 10 a =-2.7et 4.5o** 20 30 W * t for b ll 0 II 4.73 and t for b = 1 = .46 * II 0 ll * t for a .61 Figure 18. Number of racemes per plant (5 x 5, F2). 87 Number of Pods Per Plant 5 x 5, F 61 The analysis of variance shows that there were no significant differences among the replicates but there were highly significant differences among the genotypes (Table 2). BTS had the highest mean number of pods of 61.1 and strain 72-7427 had the lowest mean value of 23.6, SEA, strain 0674, and SB had mean values of 58.2, 54.5 and 31.8 respectively (Table l). Inspecting the mean values of the F hybrids, it 1 was found that 9 out of 10 hybrids had mean values exceed- ing the mid-parent values. Six of the 9 hybrids had mean values greater than that of the high parent. There was only one hybrid, i.e. from the cross 0674 x SEA, which gave a mean value lower than the low parent. In fact the reciprocal of the cross BTS x 72-7427 also gave a mean value lower than that of the low parent and, hence, there were only 2 hybrids and their reciprocals out of the 20 crosses (including reciprocals) that had mean values lower than that of the low parents. SEA and strain 0674 are very sensitive to ozone as noted earlier. Their hybrids seemed to suffer as much as the parents. Early abscision of leaves greatly affect the yield com- ponents. Number of pods per plant (X) is the first com- ponent in the sequence and was the first to be affected. The hybrids from the cross BTS x 0674 and its reciprocal 88 were superior to the others regarding total number of pods per plant (BTS x 0674 = 129.2 and 0674 x BTS = 121.5, Table 1). Having larger pod size and fewer number of pods per plant, strain 72-7427 appeared to give hybrids with smaller mean number of pods per plant when crossed with other parents. Strain 0674 and SEA which had similar phenotype did not give a heterotic effect when they were crossed. Figure 19 shows the Wr/Vr graph of the 5 x 5, F1. The regression is not significant (b = .19 t .07) as was the case for number of nodes (8 x 8, F and 5 x 5, F2) 1 and the point of intercept is not significantly above the origin (a = 119.51 i 54.49) according to the t-test. This indicates that there were genic interactions among the genes in the parental lines which controlled this trait. However, the locations of the array points representing SEA, strain 72-7427 and SB show that they contain a pre- ponderance of dominant genes. The position of strain 0674 and BTS indicates that they contain a preponderance of recessive genes. The estimate of heritability is -.10, which is equivalent to zero. The estimate of "D" is -471.05. Therefore, the estimation of other components were not at all reliable and no valid interpretation can be applied to this particular set of data. Possibly the 8 x 8, F1 or 5 x 5, F2 data can give a better picture of the gene system controlling this complex trait. Wl' 89 .1 IIIBTS .43SEA 02:88 05372-7427 .3 I'0674 800 ’ b a 19 107 * + m: 600, a =119.51 _ 54.49 1 400 * . 2 O .3 200' 5 7 ‘4 o L . #L 500 1000 1300 Vr *t for b=0 =2.65 t fOl' b I 1 =11.23 **t for a=0=2.19 Figure 19. Number of pods per plant (5 x 5, Fl). 90 8 x 8, F1 The analysis of variance indicates that there were highly significant differences among the genotypes. The mean values of strain 0685, BTS, SEA, Tui, strain 0674, Jules, SB and strain 72-7427 were 60.3, 53.6, 51.6, 50.6, 49.8, 39.9, 35.1 and 21.3 respectively (Table 3). All the 28 F1 hybrids examined had mean values greater than that of the mid-parent values. The cross BTS x 0685 and its reciprocal gave hybrids with the highest mean values of 102.8 and 92.5 respectively. The array means of BTS and strain 0685 were 67.7 and 66.6 which were the two highest array means (Table 8). From the above results it is reasonable to believe that using strain 0685 and BTS as parents would give progenies with high number of pods per plant. The regression of Wr on Vr was not significant (b = .19 i .18) (Figure 20). Genic interaction is indi- cated as the prevailing mode of gene action, although a certain amount of dominance and recessivity is not precluded. As the estimate of "F" was -202.97, this indicates that there was an excess of recessive genes among parental lines. Judging from the position of the array points none of the parents contain any preponderance of dominant genes. The position of the array point representing BTS indicates that it contains a preponderance of recessive genes. The positions of the rest of the 91 .1 SETS .5=72-7427 .2883 06:0685 03:0674 .7=TUI 200 . \Nr 01 g .2 5 8 3 100 . 7 .6 b = .19 1: .18* 04 a = 72.47 3 11.70 ** o A A - 100 200 300 Vr * t for b = 0 = 1.06 and t for b = l = 4.46 *1: t for a = 0 = 6.20 Figure 20. Number of pods per plant (8 x 8, Fl)' 92 array points seemed to indicate that they contain a balanced proportion of dominant and recessive genes. The evidence from the Wr/Vr graph indicates that genic interaction is involved in controlling this trait in the 8 x 8, F generation. 1 The heritability estimate is only 3%, a value not uncommon for traits as subject to complex genic and environmental influences as is this one. 5 x 5, F2 The analysis of variance of the F2 populations shows that there were significant differences among the replicates and among the genotypes as they were in the 8 x 8, F1. Eight out of 10 F2 populations had mean values greater than their mid-parent values (Table 5). Two of the crosses that had mean values smaller than that of the mid-parent values involved the crosses between SB x 72-7427 and SEA x 72-7427. If the array means of the 5 x 5, F1 were compared with that of the 5 x 5, F2 it can be seen that all the array means decreased in the F2 populations except those derived from strain 72-7247 (Table 7). This is consistent with effects influenced by genetic heterozygosity. The regression of Wr on Vr (Figure 21) shows that it was significantly different from zero but not signifi- cantly different from unity (b = .92 i .18). The 93 .1=BTS .438EA CZaSB .5372-7427 170 P 0330674 150 ’ 100’ b =.92 1.18 * a =27.23t4.13 H 50 tfm'bso-4J7;tfm'b'1-A5 thf a:0:1.59 50 100 150 Vr Figure 21. Number of pods per plant (5 x 5, F2). 94 regression line intersects the Wr axis above the origin (a = 27.33 i 4.13) but not significantly different from O which indicates complete dominance. SEA, strain 72-7427 and BTS appear to contain a moderate proportion of domi- nant genes. Strain 0674 seems to contain a certain level of dominant genes higher than SEA, strain 72-7427 and BTS. The shifting position of SB from the dominant side in Figure 19, to the recessive side in Figure 21 can only be attributed to the decrease in heterozygosity in the F2. The estimate of heritability was .21 which was acceptable. Number of pods per plant (X) is a complex trait and is one of the primary yield components. It is the first trait in the sequence of yield components and is the function of pods/racemes, racemes/node, nodes/branch, branches/plant. Thus, any environmental factors affect- ing the above traits affect "X" also. The estimate of "D" was 162.21 indicating that additive effects existed. The "F" estimate was -187.31 indicating that there was an excess of recessive genes among the parents. To get a reasonable prediction as to the degree of dominance and additive affects it is necessary to have the 5 x 5, F2 data excluding the variances and covariances of the other 3 additional varieties in the 1974 trial. 95 Length of Pod 8 x 8, Fl Length of pod was not included in the 1973 data. In 1974, it was decided to include this trait in the study. Three longest pods were taken from the 8 x 8, F1 and 5 x 5, F2 plant population to determine the average length of pod of each plant. The mean length of pod of the 8 x 8, F1 is shown in Table 3. Among the 8 parental lines tested, 72-7427 appeared to have the highest mean pod length (151.2 mm. Table 3) and strain 0674 had the shortest pod of 80.4 mm. Fourteen out of the 28 hybrids examined had mean values greater than that of the mid-parent values but there were only 4 means greater than the mean of the long pod parent. Those 4 hybrids derived from the crosses, SB x Jules, 0674 x SEA, SEA x Jules and 0685 x Tui. Fourteen hybrids had means smaller than that of their mid-parent values. BTS x 72-7427 produced the hybrid with the longest pod. Figure 22 presents the Wr/Vr graph of the 8 x 8, F1' The regression of Wr on Vr is significantly dif- ferent from b = 0 but not from b = l (b = .95 i .15). The intercept on the Wr axis is above the origin and is significantly different from 0 (a = 104.21 1 18.17). These two points, the unit slope and the intercept, taken together, indicate partial dominance for genes affecting pod length. This evidence is also confirmed * t for b 'k t for a Figure 22. Length of 96 CD1 ' 02" 3 ..3 u .43 .15 I .16 3 "5 .17 3 BTS 0674 SEA 72-7427 0685 TUI . 8 ' JULES * 68.951: .15 a = 104.01 '1 1s.17** 20m) Vr 6.23 and t for b = l = .36 pod in mm (8 x 8, F1). 97 by the / Hl/D value of .39. However, the consistency of the regression line here and later in the F2, the high h2 value of about 100% (h2 = 1.10) and high "D" value as mentioned earlier, strongly suggested that the genes controlling this trait are, on the average, largely additive in their action. Furthermore, the closeness of the regression line to the parabola suggests that the dominance effect is minimal. The position of the points representing parents Tui, strain 0685, Jules and SB indicate that they contain a preponderance of the partially dominant genes. Parents BTS and strain 0674 contain a preponderance of the reces- sive genes. Parent SEA seems to contain a slightly higher proportion of partially dominant than of reces- sive genes. Parent strain 72-7427 contains a balance of partially dominant and recessive genes. The position of strain 72-7427 shows that it is the outlier in this case. Strain 72-7427 was the determinate type with the longest pod as compared to the others among the 8 parental lines. 5 x 5, F2 Table 5 shows the mean values for parents and F2 populations. Only 4 out of the 10 F2 populations examined had mean values greater than the mid-parent values, and only one (0674 x SEA) had a mean exceeding the mean of the greater parent (Table 5). Six of the 98 F2 populations had means smaller than that of their mid- parent values. The F2 derived from SB x 72-7427 had the longest pod (126.9) and the shortest pod length was the one derived from 0674 x SEA (85.2). Figure 23 shows the Wr/Vr graph of 5 x 5, F The 2. graph has a similar picture to that of the 8 x 8, F The 1. regression of Wr on Vr is significantly different from b = 0 but not from b = 1 (b = .98 i .08) indicating an additive gene system for this trait exists in the F2. The "D" value of 6732.63 is in support of this assump- tion. However, the intercept on the Wr axis is above the origin and is significantly different from 0 (a = 138.49 1 5.11) indicating partial dominance. The /r§;75 value of .84 also confirms the existence of partial dominance. Nevertheless, the close relationship between the regres- sion line and the parabola suggests that dominance plays a small role in the determination of length of pod (White- house et al., 1958). The position of the points indicates that BTS contains the highest proportion of dominant genes. SEA contains a lesser level of dominant genes. Strain 0674 contains a balance of dominant and recessive genes. SB contains a preponderance of recessive genes and strain 72-7427 contains a lower level of recessive genes than SB. The "F" value of -5609.24 suggests that there is an excess of recessive genes among the parental lines. The 400 P 300 F Wr 200 1- 10C) r * t for b *1: t for a Figure 23. Length 99 '5 b 3.98 1.00* 4 / a =1aa.49 t 5.11 ** .11 BTS C>2 C 3 ‘3 0674 .4 a SEA .5 3 72 7427 100 V, 200 300 II 0 ll 11.71 and t for b = l = .24 0 = 5.29 of pod in mm (5 x 5, F2). 100 heritability estimate of 31% in the F2 generation is not high as compared to 100% in the 8 x 8, F The addi- 1. tional parents in the 8 x 8 add greatly to the additive variance for this trait and in addition the variance estimates are based on a greater number of plots than in the 5 x 5, and are therefore estimated with greater precision. Balancing the evidences in Figures 22 and 23, it is believed that an additive gene system exists and that dominant plays a small part in controlling pod length. Pod Dry Weight Per Plant 5 x 5, Fl The analysis of variance shows that there were no significant differences among the replicates but there were highly significant differences among the genotypes. Among the 5 parental lines, BTS had the highest mean total pod dry weight of 80.6 gm and strain 0674 had the lowest mean value of 35.3 gm. SB, SEA and strain 72—7427 had mean values of 54.9, 44.7 and 64.8 gm respectively (Table l). BTS, SB and strain 72—7427 were not affected by ozone injury. SEA and strain 0674 were the only 2 parents sensitive to ozone injury. As has already been discussed in the previous sections, strain 0674, in par- ticular, was very much affected by ozone injury. Even though it could recover and resumed its growth later, 101 the season was not long enough for the plant to produce photosynthate for the pods it produced. Consequently, a large portion of the pods were very light in weight and affected the pod dry weight of the plant. Examining the mean values of the hybrids (Table 1), it was found that 9 out of 10 crosses gave hybrids with mean values exceeding mid-parent values. Six out of the 9 hybrids had mean values greater than that of the high parent. The hybrid derived from the cross 0674 x SEA gave a mean value of 34.7 gm which was lower than 39.9 gm, the mid-parent value of this cross.. It was observed that when strain 0674 was crossed with SEA, the resulting Fl hybrids did not show any heterotic effect as compared to those hybrids derived from crosses between parents with contrasting phenotypes. The best parental combination which produced the hybrids with greatest mean pod dry weight was that between BTS x 0674 (BTS x 0674 = 153.42, Table l). The regression of Wr on Vr (Figure 24) is neither significantly different from b = 0 or b = 1, indicating large errors or genic interaction. As the additive genetic variance (D) has a negative value of -1262.99, the calculated values of other genetic components are not reliable. It is assumed that gene interaction plays a role in determining total pod weight. This is not surprising because total pod weight is a complex trait. 102 .Aam .m x my Em ca pecan Mom unmamz who com .vm wwsmflm .0 an. I O u a ecu a: 0.50N1 Poéupuneou «a «cruounuou a». ..> \m. , «0 OOON 009 900.. 00.: OONF ooow com 001 O. . CON 0 . com 1.. oo.m2......aa.$~ u a \ 13.“...3. an \ 1 \ 4 00¢ ..>> O 1 com a. A com VNOO " fl . ~N¢~-u~.u m 0 on u u 0 (mm H v C , mhn w. 103 It can be affected by number of pods per plant (X) and number of seeds per pod (X) and stand density. The interaction between these 2 yield components with the environment and the intra-plant competition among the components can greatly affect total pod weight. Since gene interaction is believed to play a role in this trait the assumption of "no gene interaction" does not fit this set of data. However, if we omitted BTS which was the only indeterminate type and strain 0674 which was sensitive to ozone injuries a new regression line can be drawn (dotted line). The new regression line may be significant and indicates a real relationship between Wr and Vr. The remaining three parents, namely, strain 72-7427, SB and SEA would fit the line very well. Thus, their position indicates that SB and strain 72-7427 contain a preponderance of dominant genes and SEA con- tains a preponderance of recessive genes. However, no attempt was conducted to recalculate the Wr and Vr of the three remaining parents. 8 x 8, Fl The analysis of variance indicated that there were significant variations among genotypes (Table 6). The mean pod weights of the 8 parental lines were: strain 0685 = 80.1 gm, Jules = 79.7 gm, Tui = 73.9 gm, BTS = 71.9 gm, SB = 65.4 gm, strain 72-7427 = 65.4 gm, SEA = 53.1 gm and strain 0674 = 41.6 gm (Table 3). 104 There were 2 out of the 28 F1 hybrids examined which had a mean pod weight lower than the mid-parent values. The cross between BTS x 0685 gave a hybrid with the highest mean value of 155.1 (Table 3). Again, the poorest combination appeared to be between 0674 x SEA which gave the mean value of 47.6 gm. The cross between SB x 72-7427, another pair of similar phenotype, also gave hybrids with low mean pod weight (SB x 72-7427 = 63.8 gm, 72-7427 x SB = 62.5 gm). It can be seen from Table 8 that the array means of the F1 derived from strain 0685, Jules and BTS were 107.5, 105.2, and 101.1 gms, respectively, which were the three highest array means. It is reasonable, therefore, to believe that if they were used as the parents in the crosses they would probably give hybrids with high mean pod weights. The Wr and Vr graph (Figure 25) shows that the regression is not significant (b = .43 t .20). Genic interaction is indicated. The array points were very much scattered in the graph. BTS appeared to have a strong influence in causing the deviation of the regres- sion line from a unit slope of 1. If the regression line were redrawn as shown by the dotted line all the array points except 1 and 6 would fit very well. Among the 5 parents tested in 1973, BTS gave the highest mean pod weight (Table 1). In 1974, strain 0685 gave the highest mean pod weight (Table 3). BTS and strain 0685 have ‘0 '1='BTS C> 2=38l3 .1 3='06713 0 4385A 1 40¢). 1300 . Wr 2C”). 10(1. 1 0A- -71.59 ll 0 II c. Figure 25. Pod dry weight per plant in gm 105 o 5 =72—7427 0 6'0685 .1 7 =‘TUI 0 8=JULES /'3 //.4 .5 / .6 /‘a 0‘ 30L0 403 500 800 Vr b = .43 t .20 * a = -71.58 190.88 ** 2.18, t for b = l = 2.83 .79 (8 x 8, Fl). 106 unique gene systems controlling this trait. BTS was the only indeterminate type among the 5 parents in 1973 and its phenotype was different from the other two indetermi- nate lines in 1974. Strain 0685 also had a very strange character and growth habit. When it was grown under greenhouse conditions in the winter for cross pollination, it became a vine type. This same line when grown under field conditions was a bush type with vigorous vegetative growth. It also had a very different plant type when compared with other bush lines grown in 1974. These two distinctive genotypes behaved differently from the others and might have sufficient influence on the regres- sion line to cause it to deviate from a slope of l. The computed negative value of "D" (-282.39) upset the whole picture of the graph. Gene interactions are, therefore, assumed to exist and the assumption of "no gene interaction" is again not valid in this case. 5 x 5, F2 Significant differences were found among repli- cates and among genotypes as they were with total number of pods (Table 6). BTS gave the highest mean pod dry weight of 71.1 gm. SB, strain 72-7427, SEA and 0674 gave mean values of 66.4, 62.8, 52.9 and 38.7 gm: respectively. Seven out of 10 F2 populations still had mean values greater than the mid-parent value. Three fr TI”. 107 F2 populations had mean values lower than the mid-parent values. The F2 population derived from the cross BTS x SB had the highest mean pod dry weight (84.3 gm, Table 5). The F2 population derived from the cross 0674 x BTS which gave the highest mean value in the F1 generation had the second highest mean value in the F2 (82.7 gm, Table 5). The array means (Table 8) show that the Fz's derived from BTS and SB gave the highest and second highest array means respectively (69.2, 65.5 gm). Tables 7 and 8 show that the array means were decreasing in the F2 generations due to a decrease in heterozygosity of the F2 populations. Figure 26 shows the Wr/Vr graph of the 5 x 5, F2. The regression line is significantly different from 0 but not from 1 (b = 1.06 1 .11). The line intersects the Wr axis below the origin but not significantly different from o (a = -18.12 a 3.38). The JFEE7B value is 2.75. The evidence on balance suggests that the prevailing mode of gene action is in the complete to over-dominant range. The position of the array points show that strain 72-7427 contains a preponderance of dominant genes. BTS and SB apparently contained a lower proportion of domi- nant genes than strain 72-7427. Strain 0674 appears to contain a preponderance of recessive genes and SEA con- tains a lower proportion of recessive genes than strain 0674. 11)8 .Amm .m x my Em Ga ucmHm mom unmflm3 who won .mN musmflm «v.3... b> . n 00“ an. ac. on 1 1 1 1 O p «00 oméuoua .3 . 60.1..qu 50.“ moo-oh.- .o. a on on.» u «no. u a :1 N no; u a .3 . 00' .r choc! a. . Omw hnvhnuh I m. an! N. a. (malt. are! p. . cop 109 The calculated value of F = -6.l3, indicates an excess of recessive genes among the parental lines. How- ever, the position of the array points of 3 out of the 5 parents are on the dominant side. Therefore the evi- dence for suggesting a slight excess of recessive genes is not reliable. The heritability estimate is 8% which is low. The prediction of results in future generations regarding this trait, therefore, cannot be made with accuracy. Number of Seeds Per Pod 5 x 5, Fl In the analysis of variance (Table 2), no sig- nificant differences were found among the replicates but highly significant differences were found among the geno- types. BTS gave the highest mean number of seeds per pod of 5.4. SEA, strain 0674, strain 72-7427 and SB gave mean values of 4.1, 3.8, 3.6 and 3.6 respectively. Nine out of 10 F1 hybrids examined had mean values exceeding mid-parent values. The hybrid derived from (BTS x SB gave a mean value lower than the mid-parent value (Table 1). The hybrid derived from BTS x 0674 gave the highest mean number of seeds per pod. Figure 27 presents the Wr/Vr graph of the 5 x 5, F1. The regression of Wr on Vr is significantly dif- ferent from zero but not from 1; (b = .86 i .81) it 01 'BTS .2388 03:0674 .488EA 110 05:72-7427 0.5 - 0-4 > 0.3 » .3 Wr .1 b3.86t,31 9: 0.2 1 a a , +_ 11* .5 02- O7 2 0.1» 4 o n . n L 4 (L1 0.2 0J3 0u4 0.5 Vr *t for b = 0 = 2.75 and t for b = l = .46 ** t for a = 0 = .30 Figure 27. Number of seeds per pod (5 x 5, Fl). 111 intersects slightly above the origin on the Wr axis (a = 0.02 1 .007) which is not significantly different from 0. The / Hl/D value is 0.77. On balance the evi- dence suggests partial to complete dominance. The position of the points seems to fit the regression line very well. SEA and SB appear to contain a higher pro- portion of dominant than recessive genes. BTS and strain 72-7427 contain a balance of dominant and reces- sive genes. Only strain 0674 appears to contain a high proportion of recessive genes. This agrees well with the "F" value of .16 which indicates a slight excess of dominant genes among the parental inbred lines. The heritability estimate for this trait in the 5 x 5, F is 29%. 1 8 x 8, F1 Two out of 3 indeterminate types used in the 1974 trial gave the highest mean number of seeds per pod. Tui and BTS gave mean values of 5.9 and 5.8 respectively. Jules, which is another indeterminate type, ranked 4th and gave a mean value of 4.6. Strain 0685 ranked lst among the determinate type giving a mean value of 4.6. SB gave the lowest mean value of 3.3. Twenty-five out of 28 F1 hybrids examined had mean values exceeding mid-parent values. The 3 crosses that gave hybrids with mean values lower than the mid-parent values were: BTS x Jules = 4.8, SB x Jules = 3.9 and SEA x Jules = 3.6 112 (Table 3). The hybrids resulting from BTS x Tui and its reciprocal gave the highest mean value. The analysis of variance indicated that there were significant differences both among the replicates and among the genotypes (Table 4). Figure 28 presents the Wr/Vr graph of the 8 x 8, F1. The regression line is significantly different from zero but not from a unit slope (b = .61 i .22) and it intersects the Wr axis significantly above the origin (a = .23 i .08). The / Hl/D value is 0.74. The evidence suggests again that the gene system controlling this trait is partially dominant. Since the array point representing Tui is closest to the point of intersection between the regression line and the parabola, it indicates that Tui carries a preponderance of dominant genes. Strain 0674 carries a preponderance of recessive genes. SB, SEA, strains 72-7427 and 0685 seem to carry a balance of domi- nant and recessive genes. BTS and Jules, which were the indeterminate types, behaved differently from the others. Both array points of BTS and Jules are below the regres- sion line. Jules is located close to the point of inter- section of the regression line and the parabola indicating that it contains a higher level of dominant genes than BTS which is at the center. The estimate of "hz" is 44%. The excess of reces— sive alleles was indicated by F = -.06. 113 01 = BTS 05 = 72-7427 02-58 08:0885 03:0674 O7=TUI 0.7 1' O4=SEA 08=JULEs 3 0 8 1 1. 015 0.5 r 5 2 0.. II 0.4 . 70 .1 Wr O3» .8 b=.81t .22* 0.2 L a: .232.08** 0.1 » o . . . 1 . 11 0:1 0.2 0.3 0.4 0.5 0.8 Vr *t for b = o = 2.75 and t for b = 1 = 1.73 * II 0 II *t for a 2.76 Figure 28. Number of seeds per pod (8 x 8, Fl). 114 5 x 5, F2 In the 1974 trial of the 5 x 5, F2, BTS still gave the highest mean number of seed per pod (Table 5) and SB still gave the lowest mean. The analysis of variance shows highly significant differences among the replicates and among the genotypes. Out of the 10 F2 populations examined all had the mean values exceeding the mid-parent values but only 2 crosses had mean values exceeding the mean of the high parent. The Wr/Vr graph presented in Figure 29 shows that the regression line differs significantly from slope b = 0 but not from b = 1 (b = 1.14 i .08). The regres- sion line intersects the Wr axis above the origin and significantly different from 0 (a = .18 i 13.5). The genes affecting this trait behave as though they were in the additive to partially dominant range. we can observe that the regression line is very close to the parabola. This relationship suggests also that the dominance is only partial. This observation also agrees well with the / Hl/D value of .38. The relationship of the regression line and the parabola suggests that dominance plays a small part in determining this trait. According to the position of the array points, strain 0674 seems to contain the highest level of dominant genes, BTS and strain 72-7427 contain the highest level of recessive genes and SB and SEA seem 115 0.51’ ’5 1 2 0-4- 4 3 b =1.14'."..08* a =182.13 ** 0.:‘1b \Nr 0.2.. 0 1=BTS ID 2==SB O 3:0674 0.1, O 4=SEA C 5372-7427 0 A l 41 0.11 0:2 0.3 * Vr t for b = 0 = 13.73 and t for b = l = 1.70 *7: t for a = 0 = 10.40 Figure 29. Number of seeds per pod (5 x 5, F2). 116 to have a balance of dominant and recessive genes. The "F" value of .22 suggests a slight excess of dominant genes among the inbred parents. The heritability esti- mate of 90% indicates that number of seeds per pod is highly heritable. Since the gene system is additive to partially dominant, it is reasonable to believe that the parents which gave the highest array mean would give progenies with high number of seeds per pod in the later generations. lOO-Seed Weight 5 x 5, F1 The New York kidney type of strain 72-7427 gave the highest mean 100-seed weight of 55.2 gm. SB, a large-seeded bean, gave the second highest mean value of 39.5. BTS, SEA and strain 0674 gave mean values of 17.3, 13.9 and 12.7 respectively (Table 1). Significant dif- ferences among the genotypes were found in the analysis of variance (Table 2). Out of the 10 F1 hybrids examined, there was only one, resulting from the cross BTS x SEA, which gave a mean value greater than the mid-parent value, exceed- ing the mean value of the larger parent. Two gave mean values exceeding mid-parent value but smaller than the larger parent. Seven gave mean values smaller than mid-parent values. The crosses between the large-seeded 117 parents tended to give large-seeded F hybrids but the 1 mean 100 seed weight was not necessarily greater than that of the mid-parent values. Figure 30 presents the Wr/Vr graph of the 5 x 5, F1. The regression of Wr on Vr is significantly dif- ferent from b = 0 but not from b = l (b = .87 i .14) indicating an additive gene system for this trait. The intercept on the Wr axis is significantly above the point of origin (a = 80.66 1 15.36). The calculated value of "D" = 349.32 confirmed that there are strong additive effects. The interception on the Wr axis is above the H- origin and is significantly different from 0 (a = 80.66 15.36) suggesting that the dominance, if any, is partial. This is also confirmed by the / Hl/D value of .47. How- ever, there is evidence that the gene system controlling this trait in dry bean is strongly additive. The con- sistency of the regression lines in 5 x 5, F and 8 x 8, 11 F1 and 5 x 5, F2 all strongly support the hypothesis of additivity of this trait. The position of the points indicates that strain 0674 and SEA contain a preponderance of dominant genes. Strain 72-7427 contains a preponderance of recessive genes. BTS and SB contain a lower proportion of reces- sive genes than strain 72-7427. The "F" value of 35.20 indicates that there is an excess of dominant genes among the parental lines. The 118 0 1 I 375 . 2 '8 SB .4 = SEA Q 3 = 0374 .5 3 72-7427 250 { .5 200 f b = .87 t .14 150 ' a = 80.66 ‘1’ 15.36 W7 4 100» t for baa-6,44 3 t fOl‘ b 3 1 3 094 t for a =0=5.25 50 * 0 A . - 50 150 200 250 Vr Figure 30. 100-seed weight in gm (5 x 5, F1). 119 heritability estimate of 82% suggests that loo-seed weight is highly heritable and further confirms that the gene system is essentially an additive one. 8 x 8, Fl The 1974 data confirmed 1973 data that strain 72-7427 and SB were the two highest in mean 100-seed weight (Table 3). SEA, however, in 1974, gave a slightly higher mean value than BTS. Jules gave mean value of 36.0 which was highest among the three new varieties and ranked 3rd of all the 8 varieties tested. Tui, which was about as late in maturing as strain 0685, gave the lowest mean value among the 3 new varieties (17.9, Table 3) and ranked 7th among the 8 varieties. Strain 0685 gave a slightly higher mean value than that of BTS and SEA. Strain 0674 had the lowest mean 100-seed weight (14.66, Table 3) among the 8 varieties. Out of 28 F hybrids examined, 12 had mean values 1 exceeding the mid-parent values and only 2 hybrids (0674 x Tui and SEA x Tui) gave mean values exceeding the mean of the parent with greater mean loo-seed weight (Table 3). Fifteen had mean values smaller than that of mid-parent values. It was observed that the strain 72-7427 when crossed with other parents, on the average would give hybrids with highest mean values. SB would give hybrids 120 with the second highest mean values in the same series. Strain 0674 would generally produce hybrids with the lowest mean values except in 2 series of crosses where 0685 x BTS and Tui x BTS would give the lowest mean values. The Wr/Vr graph in Figure 31 shows that the regression line is significantly different from b = 0 but not from b = l (b = .88 i .08) indicating additive effects for this trait. The intercept on the Wr axis is above the origin and significantly different from 0 (a = 48.56 i 5.25) indicating partial dominance. The closeness of the regression line to the parabola sug- gests that dominance plays a small part in the determi- nation of 100-seed weight (Whitehouse et al., 1958). As the array points of Tui, strains 0674, 0685 and Jules are closest to the lower intersection of the regression line and the parabola, they possess a preponderance of dominant genes. SEA possesses a lower proportion of dominant genes than the above mentioned 4 parents. SB and BTS contain a preponderance of recessive genes. Strain 72-7427, the outlier, has its array point below the regression line and away on the high Vr value side. Strain 72-7427, which is the red kidney type of dry bean, has the highest mean 100—seed weight. It is the only variety that behaves differently from the others. It is interesting to observe that the Wr/Vr graph of 100-seed weight, 8 x 8, F has a similar picture to 1 121 that of the length of pod 8 x 8, F (Figure 22). The l changing position of strain 0674 in Figure 22 to the position in Figure 31 is believed to be a consequence of susceptibility to ozone injuries of strain 0674. Ozone injuries did not affect length of pod but it did affect 100-seed weight. The heritability estimate of 100-seed weight is 90% suggesting that this trait is highly heritable similar to length of pod. The "D" value of 210.89 indicates the presence of substantial additive genetic variance. It is, therefore, to be expected that the parents with high array means will produce progenies with high lOO-seed weights. The correlation between length of pod and 100-seed weight of plants in the 8 x 8, F1 was calculated and had the value of 0.47 which is highly significant. The evi- dence from the correlation and Wr/Vr graphs suggested that these two traits might result from action of the same gene system. 5 x 5, F2 The ranking of the mean lOO-seed weight in the F2 (Table 5) is the same as those reported in the 8 x 8,. F1 section, namely, strain 72-7427 gave the highest mean value and strain 0674 the lowest. Four of 10 F2 popu- lations examined had mean values greater than the 122 1:81’8 . 5:72-7427 2:58 0 6=0685 330674 0 7=TUI 4: SEA 0 8=JULES 150' .1 2 0 100' 5 .4 z b =.88 t.08 * VVr ll} ’6 a = 48.58: 5.25 ** 3 7 50 7—7 50 100 Vr *t for b = 0 = 10.58 and t for b = 1 = 1.45 *1: t for a = 0 = 9.26 Figure 31. 100-seed weight in gm (8 x 8, F1). 123 mid-parent values. Six F2 populations had mean values smaller than the mid-parent values. Two out of the 4 F2 populations which had the mean exceeding the mid-parent value involved the crosses between BTS x 0674 and BTS x SEA. The other 2 F2 populations involved the cross between 0674 x SEA and SB x 72-7427. BTS x 0674, BTS x SEA and SB x 7427 also showed heterosis in the 5 x 5, F1. The Wr/Vr graph presented in Figure 32 shows that the regression line is significant (b = 1.01 1 .06) and it intersects significantly above the origin (a = 69.61 1 2.44). As far as the relationship between the regression line and the parabola is concerned, it takes the same form as in the 8 x 8, Fl. Dominance plays only a minor part in the determination of 100-seed weight. The position of the points show that BTS, strain 0674 and SEA possess a preponderance of dominant genes. Strain 72-7427 contains a preponderance of recessive genes. SB seems to contain a balance of dominant and recessive genes. The "F" value of 147.11 indicates that there is an excess of dominant genes among the parents. This agrees very well with the position of the 3 parents being closest to the point of intersection of the regression and parabola. The "D" value of 320.42 and the high value 2 of h of 81% indicate that the gene system is additive and the trait is highly heritable. Evidence from the 124 .1 BTS 02:33 04=sEA .3= 0674 .5: 72-7427 1300 1 2(H31 “H 100 ID'= 1-01 1.08* a: 606112.44 ** 100 200 300 Vr 0 = 16.24 and t for b = 1 = -.11 'k t for b 19* t for a 0 - 11.65 Figure 32. loo—seed weight in gm (5 x 5, F2). 125 three graphs strongly suggests that the gene system con- trolling 100-seed weight in dry bean is additive and the trait is highly heritable. Number of Seeds Per Plant 5 x 5, Fl BTS had the highest mean number of seeds per plant (338.9, Table l) and strain 72-7427 had the lowest mean of 85.6. SEA, strain 0674 and SB had mean values of 235.9, 201.7 and 109.7, respectively. Eight out of 10 F hybrids I examined had mean values exceeding that of mid-parent values and 2 hybrids had the mean values smaller than mid- parent values. The hybrid with the highest mean value derived from BTS x 0674 which gave a mean of 720.9 (Table l). The lowest mean value came from the hybrid of 72-7427 x BTS which gave a mean value of 62.3. It was disturbing that the computed "D“ value had a negative value of -1l798.58. The high error variance (E = 23580.39) coupled with the low variance of the parental arrays resulted in a very high negative "D" value. The regression is not significantly different from 0 but significantly different from 1 (b = .18 1 .08) (Figure 33). The line intersects above the origin but not significantly different from 0 (a = 4532.73 1 8098.79). Further interpretation is not valid. 126 .Aam .m x my woman mom mommm mo Hmnfisz .mm musmflm mH.N n o u m MOM #4 k. .> oo.ea u a n a how u .ma.m u o u n how 01 ooodv ooodn 03.3 03.3 1 . . 1 1 o 1.. Summon u 2.3: u u no 3% Q0. H ”—1. n n 1.0 «0 F0 . 30.3 no #506" n . . OO0.0N 3232.160 mun «0 Sauce meuuao 1 . . :s 127 Number of seeds per plant is a complex trait. Several factors can influence the variation in this trait. It is strongly influenced by number of pods per plant (X) and number of seeds per pod (Y). In turn, "X" and "Y" themselves are strongly influenced by environmental factors. Inter-plant competition can easily cause variation in "X." Once the pods are formed there may be intraplant competition if environmental stress is imposed. Although "X“ may not have any direct genetic influence on "Y," nevertheless, developmentally it has an indirect effect on “Y" (Adams, 1967). The interactions between X and Y and between these two components with the environment may cause "D" to be negative. It is, there- fore, not reliable to make further interpretation of this set of data. 8 x 8, Fl In the 1974 trial, BTS still showed its superi- ority in this trait over the other 7 lines. It had a mean value of 318.9, as compared to the lowest mean value of 82.2 for 72-7427 (Table 3). Twenty-three out of 28 F1 hybrids examined showed heterosis, having mean values greater than that of mid-parent values. Only 5 hybrids had mean values smaller than that of mid-parent values. The F1 hybrid derived from BTS x 0685 gave the highest mean value of 662.2; the smallest mean came from the hybrid SB x 72-7427 (103.5). 128 Figure 34 presents the picture of the Wr/Vr graph of 8 x 8, F1’ The regression coefficient has the value of .24 1 .05. This is significantly different from b = 0 and b = 1. Genic interaction is indicated for this trait. The intercept on the Wr axis is above the origin (a = 3377.79 1 580.07) and it is significantly different from 0. Strain 72-7427, Jules, SB, Tui and SEA seem to contain a preponderance of dominant genes. BTS contains a preponderance of recessive genes. Strains 0674 and 0685 seem to have a balance of dominant and recessive genes. Even though the calculated value of "D" is 2968.64 which indicates additivity of the gene system, genic interaction should be postulated for this trait according to the evidence of the regression line. The heritability is low (h2 = 5%). 5 x 5, F2 Seven out of 10 F2 populations examined had mean values greater than that of the mid-parent values (Table 5). The F2 population derived from the cross BTS x SEA gave the highest mean value of 303.5. Three of the Fz's gave mean values lower than that of mid- parent values and the lowest mean value belongs to the cross derived from SB x 72-7427 (92.2, Table 5). The regression of Wr on Vr (Figure 35) is sig- nificantly different from b = 0 but not from b = 1 129 .Aam .m x mo Madam mom momom mo Hmnasz .vm musmflm ..> OOOMN OOOON ooomw OOOO— DOWN 1‘ 11 d m. Nomuonabnza h \m.\ 000v . . . 1 .. \ oomwnwuneo.~.mmctonntouu N. \ :5 F0 I n h0.00m+ah.hhmn n— a no. H ea. u a . cocoa .OOOMP ehoc "a. nLuVN11Nh nu 0 av mm_nunu .. L OOOwa 5. com 00¢ com CON .0 1 1 1 o a. F. 1. ohmopuameo n a L or «spinach... «0 m0 0 0. .>> 0 c. n. . OON mwaaaflo. (meV. Shun. chooun. mooonc. mm "N. hNVh-Nh 1.1.. m. who .1- w. 1 000 138 SEA x 0674 (31.1). The analysis of variance (Table 6) .showed that there were highly significant differences among replicates and among genotypes. Figure 38 shows that the regression of Wr on Vr is significantly different from b = 0 but not from b = l (b = .99 1 .12). The regression line intersects the Wr axis below the origin but not significantly different from 0 (a = -9.09 1 3.06) indicating complete dominance. This result does not agree with the J Hl/D value of 2.41 which indicates over-dominance. On balance the evidence from the graph and the J Hl/D value, suggests complete to over-dominance. The position of the points indicate that strain 0674 and SEA contain a preponderance of recessive genes. Strain 72-7427 contains a preponderance of dominant genes. SB and BTS contain a lower proportion of dominant genes than strain 72-7427. The calculated "F" value of -28.67 indicates that there is an excess of recessive genes among the parental lines. The heritability estimate of 10% suggests that total seed weight is of low herita- bility. It is not expected that total seed weight or seed yield (W) per se will be transferred to the next generation effectively. It has been understood that "yield" is a complex trait. "W" is very much influenced by other components i.e. X, Y and Z. Any factor affecting these components would ultimately effect "W." 120 100‘ 50 139 012875 04-5911 0 2'53 05:72-7427 0 330874 991.12 * ink - 90918.06 100 120 Vr * II 0 II 8.28 and t for b = l = .06 t for b ** II C II t for a .97 Figure 38. Seed dry weight per plant, in gm (5 x 5, F2). 140 Harvest Index 5 x 5, Fl The variety SB had the highest mean harvest index among the S parental lines tested in 1973. Its mean value was .63. Strain 0674 had the lowest mean harvest index of .50 (Table 1). Eight out of 10 F hybrids 1 examined had the mean harvest index greater than that of the mid-parent values and two smaller. BTS x SB pro- duced the hybrid with the highest mean value of .64 and SB x 0674 gave the hybrid with the smallest mean value of .50. The analysis of variance shows that there were significant differences among the genotypes. Due to the very small value of variances and covariances it was difficult to construct a Wr/Vr graph for this trait. However, according to the calculation, it was found that the regression coefficient had the value of b = .96 i .15 and it was not significantly different from b 1 but significantly different from b = O. This suggests that either an additive or domi- nant gene system, without the complication of genic interactions, controls this trait. The "P" value of -.0026 indicated that there was slightly an excess of recessive genes among the parental lines. The heritability estimate was 7%. This indicates that harvest index pro- bably is not highly heritable. It is a complex trait and depends on several other factors. The plant dry weight 141 or the biological yield of the plant which is the denomi- nator is heavily dependent on other traits and environ— mental factors. The total seed yield which is the numerator for harvest index is a complex trait by itself and influenced by the interactions of the yield compo- nents (X, Y and Z). Harvest index is a ratio of seed yield to total plant weight and thus expresses the effi- ciency of partition. As such it is useful in selection, but it does not by itself indicate the highest yielding lines. 8 x 8, Fl Table 3 shows the mean H.I. of the 8 x 8, Fl' SB still had the highest mean value of .66 among the 8 parental lines tested in 1974. Strain 0685 appears to have the lowest mean value of .44. Seventeen out of 28 crosses gave F1 hybrids with mean values greater than that of the mid-parent values and 11 F hybrids had mean 1 values smaller than that of the mid-parent values. . SB x Jules appeared to give the hybrid with the highest mean value of .70 and the lowest mean value came from the hybrid derived from SB x 0685 which had the value of .45. It is shown in Table 8 that Jules had the highest array mean of .65 and strain 72-7427 the lowest of .51. 142 SB and SEA had the next highest array means of .59. The lowest array mean belongs to the strain 72-7427 (.51). S x 5, F2 Eight out of 10 F2 crosses had the mean harvest index smaller than that of the mid-parent values. The F2 derived from BTS x 0674 and 0674 x 72-7427 had the smallest mean value of .52 (Table 5). Only 2 F2 crosses had mean values greater than that of the mid-parent values. They were those derived from SB x 72-7427 and 0674 x SEA which had mean values of .61 and .56 respec- tively. However, these 2 crosses appeared to give the osmallest mean seed yield (43.4 and 39.3, Table 5). On the other hand, BTS x SB, which gave a H.I. of .60 and is lower than the mid-parent value, had the highest mean seed yield of 66.8 (Table 5). Therefore, harvest index is not a reliable trait in predicting the yield. DISCUSSION At the outset of this thesis it had not been determined which morphological traits were most important in constructing an ideotype for high yield. Consequently a rather large group of characteristics had to be included in the analysis. Denis (1971) did a factor analysis of plant-type variables related to yield of dry beans. A Weight-factor and a Number-factor, typified by pod weight and number of pods, respectively, identified as two major factors or patterns for high yielding bean plant type. A third factor, a Display-factor, was inferred, based on upper internode length and leaf size. The genetic complexity of the traits proposed for investigation here have not been thoroughly studied. A quick assessment of certain aspects of the traits, e.g. gene action, etc., is necessary. One would want to know when he is building an ideotype, how the gene system of these traits will interact or recombine. The diallel cross was thought to be an appropriate method for this investigation. Generally we can get reliable information in a rather short time. While we are using the parental materials making the cross for genetic studies, we also 143 144 generate populations for selection of potential new varieties. These new varieties may provide details in genetic analysis. In order to assure that a wider base of genetic differences would be expressed for these char- acteristics among parental lines involved, the size of the diallel in this study was extended from 5 in 1973 to 8 in 1974. There are limitations in the diallel methods. Firstly, it is understood that in dealing with single plant data, large errors of estimation are expected. Secondly, with a small number of parents in a diallel set, one extreme parent can have a disproportionate influence on the slope of the regression line. This will lead to genetic interpretation that are valid only for that particular set of parents (5 x 5) and may not be valid for another set of 8 parents. In this study it can be seen in Figures 8, 14, 24, and 36 that parent 1 (BTS) behaved as an outlier in the 5 x 5, either in the F1 or F2 generations. This is because BTS has a con— trasting plant type and, presumably, genotype as compared with others in the same set of parents. In some cases, the strain 0685 joined BTS as the outliers. This can be seen in Figures 13 and 25. Here again, even though strain 0685 is a determinate type, its morphological characters are much different from other determinate types among the 8 parental lines. Therefore, both BTS and strain 0685 145 behaved differently from the others in some of the traits. They tend to have low covariances and high variances. This may have caused the slope of the regression line to deviate significantly from 1, thus influencing the inter— pretation to be genic interaction. For the purpose of building an ideotype, we must have appropriate genetic recombinations. One can obtain preliminary information about recombination in the F2 generation by correlation of the traits under investi- gation. The degree of association of the traits in the F2 has to be observed. High heritability was indicated for some traits, namely; number of seeds per pod and number of seeds per plant in the 5 x 5, F2, loo-seed weight in all 3 sets of data, and length of pod in the 8 x 8, F This can 1. be used partly as evidence of additivity of gene action. In several cases, however, low h2 values were indicated together with evidence of genic interaction. In the case of days-to-first-flower, for example, in 5 x 5, F1 and 8 x 8, F apparent partial to complete dominance was 1: indicated. In the 5 x 5, F2, however, genic interaction was indicated. Genic interaction as deduced from the Wr/Vr graph, would obscure the manifestation of genic additivity or dominance if these were present. It does not, in itself, exclude them. These kinds of gene action may, in fact, be operating in the expression of this 146 trait, but simultaneous presence in two or more parents of genes that interact in a nonadditive way with each ' other, leads to regression slope less than unity. And it is on the slope evidence that genic interaction is inferred. Often, it will turn out that, by omitting one or possibly two arrays from consideration in the makeup of the Wr/Vr graph, the slope value will return to unity, and the genetic interpretation of simple additivity or dominance will be indicated (Dickson, 1967). Therefore, the deduction of genic interaction in some of the traits in this investigation does not mean that additivity or dominance are thereby excluded. Some of the more important characteristics we want to investigate, e.g., are size and number character- istics. These traits seemed to be more complex in their genic behavior. In some cases they showed additivity such as length of pod, number of seeds per pod (5 x 5, F2) and loo-seed weight. In other cases, they show genic interaction such as length of internode (8 x 8, Fl), number of seeds per plant and seed weight per plant. One of the most important components of yield, number of pods per plant, showed genic interaction in the 5 x 5, F 1 and 8 x 8, Fl' In the 5 x 5, F2 complete dominance was indicated. It can be seen that some of the important traits are not simple genetically.‘ However, number of seeds per pod and lOO-seed weight were found to have high 147 heritability and their gene system is additive. It would be extremely lucky for a breeder to find that all the major traits contribute to high yield are highly heritable and also have additive gene systems. One would generally find that one of these traits have a low h2 value and genic interaction may be involved. It is also very com- mon to find that these three important traits, namely; number of pods per plant, number of seeds per pod and loo—seed weight, usually have negative correlations (Adams, 1967). This, of course, makes it more difficult for plant breeders to create an ideal plant type for high yield. Overall, it has not been found very promising in this investigation for all the traits investigated. This is due to the fact that most of the traits do not have an additive gene system. Genic interaction seemed to be a major feature of the genetic behavior of several traits. It appeared likely that this was caused, in part, by the inclusion in the diallel of particular parents (e.g. BTS). In any case, genic interaction does not exclude additivity or dominance. The final test would be to conduct multiple- trait selection experiments in advanced generations, wherein genic heterozygosity would be minimal and where reliable mean values could be obtained. S UMMARY AND CONCLUS IONS Inheritance of morphological characteristics of field beans (Phaseolus vulgaris L.) was studied in (l) a 5 x 5, F and F2 diallel, and (2) an 8 x 8, F diallel. l l The parental materials were hand pollinated in the green— house in the winter of 1972 and 1973 to produce the necessary crosses. The parental and F1 seeds of the 5 x 5 diallel were space planted in the field in East Lansing in the summer of 1973. The parental, P1 of the 8 x 8 and F2 of the 5 x 5 were space planted in the adjacent field in the summer of 1974. Data collection, either on living plants in the field or subsequent determination after harvesting was done on a per plant basis. Some of the missing data were replaced by figures calculated according to Snedecor and Cochran's (1976) method. For the late maturing plants which some of whose pods had been damaged by frost yield components (X, Y and Z) were determined from the fully matured pods of that plant. The results of this investigation are summarized in the following paragraphs. 148 149 1. An additive gene system was not found for the trait number-of-days-to-first-flower for the 5 x 5, F1 and an apparent partial to complete dominance was indi- cated for the 8 x 8, F Even though the evidence from 1’ the 5 x 5, F2 graph indicated genic interaction for this trait, this does not mean that additivity or dominance were excluded from the gene system. The majority of the F1 hybrids showed heterosis. The cross involving two determinate parents, SB x 72-7427, gave hybrids with the earliest flowering date and one indeterminate (BTS) x determinate (0685) cross gave hybrids with the latest flowering date. The heritability was found to be low (29%). 2. BTS was found to have the shortest duration- of-flowering among the 5 varieties tested in 1973. The longest duration-of-flowering of strain 0674 appeared to be abnormal due to ozone injury in an early growth stage which induced a resumption of growth and reproduction in a delayed later stage. The heritability estimate was 15%. Genic interaction was involved in controlling this trait. 3. The earliest maturing variety was SEA with a mean of 74.77 days. Strain 0674 with a mean of 92.63 days was found to be the latest in maturity. However, the delay in maturity of strain 0674 was probably due to ozone injury and the resumption of its growth later in 150 the season. Genic interaction was found in the 5 x 5, Fl' In the 8 x 8, F1 multiple alleles was found to control this trait. Whereas in the 5 x 5, F2 complete dominance was indicated. It could be hypothesized from the F2 data that late maturity is controlled largely by dominant and early maturity by recessive genes. Heritability was found to be low in all three sets of data (7%, 29%, 14%). 4. BTS and strain 0685 had the highest mean plant weight in the 5 x 5, and 8 x 8 diallels respectively. Strain 0674 had the lowest mean plant weight in both years. It was noticed that parents with similar plant type, e.g. strain 0674 and SEA, when crossed, gave hybrids with poor mean plant weight. On the other hand, crosses between parents with contrasting plant types, e.g. BTS x 0674 or BTS x 0685, gave hybrids with largest mean plant weights (Table l and 3). Genic interactions as well as genic dominance effects were believed to be involved in determining this trait. The heritability was very low (4%). 5. Strain 0674 produced the highest number of main stem branches among the 5 parents tested in 1973 and SB the lowest. Eight out of 10 hybrids in the 5 x 5 diallel showed heterosis. However, when parents of similar phenotype, e.g., 0674 and SEA, were crossed, the hybrid did not show heterotic effects. This trait showed great response to plant density, and to differential 151 vigor within plots attributable to legitimate genetic segregation (in the F2). Consequently, the errors associated with its estimation were large, tending seriously to bias estimates of any genetic effects that might have been present. The estimate of "D," in the 5 x 5, F1 was in fact negative. Accordingly, no attempt was made in 1974 to measure number of branches. 6. BTS had the highest mean number of main stem nodes among the 5 parents tested in 1973. Seven out of 10 F1 hybrids showed heterosis for this trait. All the parents except BTS appeared to contain a preponderance of recessive genes. The estimate of F = -24.36 seemed to agree with this interpretation. However, the estimate of "D" was negative (-l7.24) and it was decided that no valid interpretation could be reached based upon the 5 x 5, F data of 1973. l 7. Strain 0685 gave the highest mean total number of nodes among the 8 parents tested in 1974. BTS x 0685 gave the hybrid with the highest mean value of 196.50. Eight out of 10 F2 populations from the 5 x 5 had mean values exceeding the mid-parent values. Genic inter— actions were indicated for this trait in the 8 x 8, Fl. In the 5 x 5, F2, genic interaction was indicated as the prevailing mode of genic action when heterozygosity was 152 reduced to that of the F2 level. An excess of recessive genes among the parents was also suggested (F = -24.36). The heritability for this trait was very low (3-4%). 8. The determinate type appeared to have average internode lengths longer than the indeterminate type. Strain 0685 had the longest internode length. Seventeen of 28 hybrids showed heterotic effects in the 8 x 8, F1' The hybrid with the longest internode length was derived from 0685 x SB and the shortest was from Tui x Jules. Genic interaction was suggested for length of internode in the 8 x 8, F The F2 population with the longest 1. internode was derived from 0674 x 72-7427 and the F2 with the shortest internode resulted from BTS x 0674. A partial to complete dominance was suggested for length of internode in the 5 x 5, F2. Since all the indetermi- nate varieties had array points near the origin in both the 8 x 8, F and 5 x 5, F2 graphs it can be hypothesized 1 that short internodes are controlled by dominant genes. The positions of the determinate varieties on the graphs also suggested that long internodes were controlled by recessive genes. Heritability was 22% in the 8 x 8, F1 and 14% in the F2. 9. Strain 0685 was found to have the highest number of racemes among the 8 parental lines. The hybrid derived from BTS x 0685 gave the highest number of 153 rwacemes. Additive genetic variance was not detected. The Wr/Vr graph indicated genic interaction. However, if 0674, 0685 and BTS were omitted and a new regression ILine drawn, the remaining array points would fit the regression line quite well, implying for the remaining Iparental arrays an additive to partially dominant system. liine out of 10 F2 populations retained heterosis for Iaumber of racemes. The F2 with the highest mean value \was derived from BTS x SEA. The evidence from the graph and W in the 5 x 5, F2 is strongly suggestive of «complete to slight over-dominance. All the parents seemed to contain a higher proportion of recessive genes. 'The negative value of "F" supported this observation. The heritability estimate in the F2 for this trait was 11%. 10. BTS and 0685 had the highest mean number of pods per plant (X) in the 5 x 5 and 8 x 8 diallels respectively. The only hybrid in the 5 x 5, F1, 0674 x SEA which did not show a heterotic effect was probably adversely influenced by ozone injury. BTS when crossed with 0674 in the 1973 and with 0685 in 1974 gave the hybrids with largest mean number of pods per plant. The parent with large pod size, 72-7427, when crossed with the others, tended to give hybrids with small 154 number of pods per plant. Genic interaction was found both in the 5 x 5, F and 8 x 8, F Complete dominance l 1' 'was found to control number of pods per plant in the F2. An excess of recessive genes among the parents was indi- cated for all p0pu1ations tested. The heritability esti- mates were negligibly low in 5 x 5, F and 8 x 8, F 1 1 ‘whereas it was 21% in the 5 x 5, F2. 11. Strain 72-7427 had the longest pod length among the 8 parents. BTS x 7427 gave the hybrid with the longest pod in 8 x 8, F but in the F2, the family 1 derived from SB x 72-7427 had the longest pod. Additivity was indicated in both 8 x 8, F and 5 x 5, F2. There was 1 an excess of dominant genes in the 8 x 8, F1 whereas an excess of recessive genes was indicated in the 5 x 5, F2. Heritability was 100% in 8 x 8, F and 31% in the F l 2' The closeness of the regression line to the parabola in both graphs indicated that dominance plays only a minor role in determining length of pod. 12. As in number of pods per plant, BTA and strain 0685 had the highest mean pod dry weight in 1973 and 1974, respectively. The hybrids with highest mean pod dry weight were those derived from BTS x 0674 in 1973, and 0685 x BTS in 1974. These hybrids were the results of crosses between parents with contrasting plant types. Both the 5 x 5, and 8 x 8, F 1 gave a 155 negative value of "D" which follows from the large error variances associated with estimates of pod dry weight, and the low estimates of additive genetic variance associated with differences among array means. The Wr/Vr graph suggested that the genetic variance observed ‘ was generated mostly by genic interaction. However, in the F2, complete to over-dominance was indicated. The F2 generation with the highest mean pod dry weight was derived from BTS x SB. The heritability was low (8%) . A slight excess of recessive genes among the common parents in the F2 was also suggested. 13. BTS and Tui gave the highest means number of seeds per pod in 1973 and 1974, respectively. BTS x 0674 gave the hybrid with highest mean number of seeds per pod in 5 x 5, F1 and BTS x Tui gave the highest in 8 x 8, F Partial to complete dominance was indicated 1. in the 5 x 5, F Partial dominance was indicated in 1' the 8 x 8, F The evidence from the graph suggested 1. that dominance played a minor role in the 5 x 5, F2. It followed that an additive gene system was suggested. Heritability was 29% in the 5 x 5, F1, 44% in the 8 x 8, F1 and 90% in the 5 x5, F 2. l4. Strain 72-7427 had the highest mean lOO-seed weight among all parents tested. The hybrid derived from SB x 72-7427 had the highest mean lOO-seed weight in k>oth years. The Wr/Vr graphs of the 5 x 5, Fl' 8 x 8, 156 I} and 5 x 5, F2 indicated that the gene system con- trolling lOO-seed weight is highly additive which agrees *well with the positive value of "D" in all 3 graphs. ‘The heritability was high, ranging from 82% for 5 x 5, F1, 90% for 8 x 8, F and 81% for 5 x 5, F On balance, 1 2' the evidence from all 3 sets of data strongly indicates additivity with minimal dominance effect. It was in- ferred from the ”F" value that there were an excess of dominant genes in all 3 sets of data. The correlation test showed that loo-seed weight was significantly cor- related with length of pod (r = .4663, df = 54) and they jprobably have the same gene system. 15. BTS had the highest mean number of seeds per jplant. The hybrid with the highest mean number of seeds jper plant in the 5 x 5, F was derived from BTS x 0674. 1 In the 8 x 8, F the hybrid derived from BTS x 0685 gave 1! the highest number of seeds per plant. The F2 generation derived from BTS x SEA gave the highest mean value of this trait. No additive genetic variance was detected for this trait. Genic interaction was indicated in the 5 x 5 and 8 x 8, F Partial dominance was suggested 1. for the 5 x 5, F2. 16. BTS gave the highest mean seed dry weight in the 5 x 5, F and Jules gave the highest mean in the 1 8 x.8, F The hybrid derived from BTS x 0674 gave the 1. highest seed yield in 1973 and the hybrid derived from 157 Jules x 0685 gave the highest in 1974. Additive variance was not found for this trait. Gene interaction is believed to be involved but the evidence is not com- pelling. In the 5 x 5, F2, however, partial dominance was indicated. The F2 derived from BTS x SB had the highest seed yield. An excess of recessive genes among this set of parents was indicated in F2. The herita- was negligible bility in the 5 x 5, F and 8 x 8, F l 1 Jbecause of negative ”D" values, whereas it was 10% in the F2. 17. BTS was found to be a promising variety to Zbe used in a breeding program for improving seed yield. frui may be an alternate choice. APPENDIX nu nu mo.aem mm.HHHH mm.mme ma.ammn mm.H~eu psmam\.uz mun tmmm .. u- am.ommm~ mm.amaa~ sm.emee~ Hm.ameemu ~m.maeaeu usdad\mnmmm * mm.o ee.o am.m mm.mm ee.ee o~.mm ~m.aem .pz tmmmuooa am.o ee.o e~.o a~.o a~.e ea.e om.o Awe dod\memmm * u- .. aa.eeme e~.HeeH mm.~maa ~e.mmeau mm.memeu ssdHE\.uz and mod n. .. om.m~m em.~ee eo.eem e~.eaaau me.Heeu Axe usmad\me0d a u- .. He.aa Hm.m- e~.oau mm.e~u em.eau mmeoz smum sad: 4 u- .. om.m em.m No.e mm.Hn e~.- mwsosmsm swam sees * u- .. ae.eeam we.eeem mm.mmee Hm.~meen em.mmmmu .uz who added so. o~.~ ee.ae mo.ooa ~0.HHH HH.OH1 ea.- spanned: 0» memo mH.e mm.a mm.a am.m~ Ho.e~ aH.m ee.oe asahmroee mo soaumsso Ha.o ee.a we.m Ho.ee we.ee me.amn -.ma heroes use op when as .mmmmx m mm am e o .Aam .m x my mGOHHMHHm> mo ucmcomfioo owumcmw . m wands 158 159 u: II «m.mvm mm.mhb mm.mm> mm.mvvl mo.em| RSV “swam\.uz mun comm mo.o HB.N mm.a¢m¢ mH.NmHHN Hm.oawam mh.mmmwa| vm.mom~ unmam\m000m # om.o mv.o mo.v mN.hm vm.mv om.mm mm.0HN ANV .uB vmmmlooa ve.o vh.o mH.o nm.o mv.o mo.o1 mb.o hwy vom\mommm # II II m~.mmv mo.hmma ma.mmma Hm.NNmI mm.~mm| unmam\.u3 mum pom oa.H mm.o Ho.oa mm.mm mm.mn «v.mma hm.mhv pom mo sumcmq mo.o oe.m hh.v~H mo.mmv mo.mmv nm.moml Hm.bm “xv usuam\moom * -u .. mo.mm ee.mea me.mma ma.omu mm.~u usdae\mmsmomm * mm.o HN.H mm.~m~ vm.~mva mm.NHhH mm.mmmH oe.mnaa mooshmucH mo zumcwq mo.o mN.v mv.mam mh.mvm~ nm.mmv~ mv.o~nl av.mma wmwoz mo Hmnfisz Hmuoa II II oa.mhm 5H.Nmam om.mm~m oa.mhmal mm.mmm| .uz mun unwam m~.o mH.H >N.ma ma.om ma.mh mm.oml mm.Hm unassumz op mama ~m.o HH.H mm.v om.mm Nv.hm mm.oml mm.mv HmSOHm uma on whom an .o|\.H|m\ m mm Hm m o .Aam .m x my mcofl»0flum> m0 usmcomsoo owumcmo .oa magma 160 oa.o He.~ ee.am He.mam em.ome em.m~n me.ee 13V usde\.p3 who tmmm oa.o mm.o me.am mH.ooH em.ee~ Ho.m~m ma.eme usde\memmm * Hm.o ~m.o em.~. mo.mma mm.eH~ Ha.eea Ne.o~m law .93 emmmuooa oa.o mm.o mo.e oe.o NH.o -.e em.o Awe todxmemmm # mo.o me.m Ho.aa e~.ea~ m~.Hee ma.mu em.ooe usmad\.uz who toe Hm.o em.o ae.eeaa am.eeme mw.meea e~.aoemn me.~mem was no rumsmq H~.o mm.a mo.He ea.me~ «m.mow Hm.emau H~.~mH Axe usmam\mooe # HH.o ea.~ he.m mm.~oa om.aoa mm.amu ea.m~ usdam\mmsmomm # ea.o ma.H mo.Ho~ oa.mea~ e~.eae~ we.mveu ea.mme stoneware mo spasms eo.o Ho.e Ho.mm em.emo~ me.HH~H mH.eHeu om.ma mmtoz mo amnesz Hmuoa eo.o mo.e mm.eee em.~ema me.emma oe.ma~n em.ma .93 and named eH.o mm.~ Hm.o He.ee~ em.ae~ ~e.me mm.~e spanned: on when mH.o ~m.a mm.~fl m~.mm om.ae mm.emu oe.am shades has on hand ms .mmmmx m «m Hm e o .Amm .m x my mGOHUMflHm> mo usmsomfioo owumcww .HH magma 161 Tade 342. Days to flowering: array covariances, variances, and their differences (5x5, F1). Wr Vr Wr—Vr 12.8048 23.4551 -10.6503 18.8084 18.5549 .2535 4.6747 5.1593 -.4846 25.3686 35.2405 -9.8719 16.3206 17.5254 -1.2048 Total 77.9770 99.9352 -21.9582 Necessary statistics for Wr Vr 22.962 23.455 20.423 18.555 10.769 5.159 28.145 35.241 19.848 17.525 plotting limiting parabola. Y? ‘ (variance of the parents) - 22.479 Br ‘ (the mean covariance of the parents and the arrays) = 15.595 :5 ‘ (the mean variance of the arrays) = 19.987 r- (the variance of the means of the arrays) = 11.005 162 Table 13. Days to flowering: array covariances, variances and their differences (8x8, F1). Wr Vr Wr-Vr 37.2017 36.3169 0.8848 51.3766 50.5102 0.8663 28.6434 22.6990 6.0444 39.2315 34.5062 4.7253 51.7221 51.3769 0.3452 26.5031 28.9868 -2.4837 16.8240 12.9861 3.8379 15.1304 8.9996 6.1308 Total 266.6328 246.2818 20.3510 Wr Vr 48.312 36.317 56.976 50.510 38.111 22.599 47.092 34.506 57.467 51.377 43.162 28.987 28.890 12.986 24.050 9.000 Necessary statistics for plotting limiting parabola. (variance of the parents) - 64.269 (the mean covariance of the parents and the arrays) = 33-329 - (the mean variance of the arrays) - 30.785 - (the variance of the means of the arrays) - 18.192 4 41:14 am r1 '0 I Table 14. Total Days to flowering: 163 and their differences (5x5, F2). array covariances, variances, Wr Vr Wr-Vr 15.1306 14.9696 .1610 23.2061 18.4337 4.7725 15.8980 16.3349 -.4369 90.5453 91.4118 -.4369 Necessary statistics for plotting Wr Vr 26.235 20.496 22.421 14.970 26.668 21.178 24.880 18.434 23.421 16.335 Vp - (variance of the parents) - 33.580 Hr - (the mean covariance of the parents and the arrays) = 18.109 V5 - (the mean variance of the arrays) - 18.282 Vr - (the variance of the means of the arrays) - 11.091 limiting parabola. 164 qhbhgls. Duration of flowering: array covariances, variances, and their differences (5x5, Fl). Wr Vr Wr-Vr 4.2163 2.9153 1.3010 9.6751 14.1254 -4.4503 3.4009 19.1637 -15.7628 2.2319 3.8662 -1.6342 Total 16.4827 49.3992 -32.9166 Necessary statistics for plotting limiting parabola. Wr Vr 4.828 2.915 10.627 14.125 12.378 19.164 8.636 9.329 5.560 3.866 (variance of the parents) - 7.995 Wr - (the mean covariance of the parents and the arrays) = 3.297 (the mean variance of the arrays) ' 9.880 (the variance of the means of the arrays) a 2.159 < 'd l < <2 '1 H l I Table 16. Days to maturity: 165 array covariances, their differences (5x5, F1). Wr Vr Wr-Vr .5119 49.0775 -48.5656 19.6605 15.2323 4.4282 37.4017 54.8509 -17.4492 40.9315 69.6673 -28.7358 15.2037 6.1551 9.0486 Total 113.7093 194.9831 -81.2738 variances and Necessary statistics for plotting limiting parabola. Wr Vr 48.825 49.077 27.201 15.232 51.617 54.851 58.172 69.667 17.291 6.155 (variance of the parents) - 48.573 (the mean covariance of the parents and the arrays) (the mean variance of the arrays) = 38.997 (the varianceof the means of the arrays) = 14.506 22.742 Table 17. Total <<213< H uric I II I Days to maturity: their differences (8x8, F1). 166 array covariances, variances and Wr Vr Wr-Vr 12.7977 39.1711 -26.3735 47.9942 72.7896 -24.7953 63.6142 78.6827 -15.0685 52.6734 48.9027 3.7707 34.7439 37.1858 -2.4419 37.4036 35.6818 1.7217 24.3398 14.0462 10.2935 30.2754 24.2070 6.0685 303.8422 350.6669 -46.8248 Necessary statistics for plotting Wr Vr 53.546 39.171 72.992 72.790 75.889 78.683 59.829 48.903 52.171 37.186 51.105 35.682 32.064 14.046 42.093 24.207 (Variance of the parents) 8 73.195 (the mean covariance of the parents and the arrays) (the mean variance of the arrays) - 43.833 (the variance of the means of the arrays) = 20.746 limiting parabola. 37.980 167 ThbhalB. Days to maturity: array covariances, variances and their differences (5x5, F2). Wr Vr Wr—Vr .9034 11.1937 -10.2903 1.9365 9.1586 -11.0951 38.8037 33.2354 5.5683 -4.1837 3.8818 -8.0655 Total 82.4259 112.0654 -29.6394 Necessary statistics for plotting limiting parabola. Wr Vr 22.824 11.194 20.645 9.159 50.407 54.596 39.328 33.235 13.441 3.882 Yp - (variance of the parents) - 46.539 Hr - (the mean covariance of the parents and the arrays) = 16.485 V5 - (the mean variance of the arrays) - 22.413 Vr - (the variance of the means of the arrays) = 6.876 168 Table 19, Plant dry weight: array covariances, variances, and their differences (5x5, F1). Wr Vr Wr-Vr —713.3178 2287.1513 -3000.4691 -240.9623 628.3488 -869.3111 1147.0317 3727.0787 -2580.0471 791.8640 1833.7598 -1041.8958 -265.3427 261.9819 -527.3246 Total 719.2728 8738.3204 -8019.0476 Necessary statistics for plotting limiting parabola. Wr Vr 950.938 2287.151 498.431 628.349 1213.918 3727.079 851.483 1833.760 321.840 261.982 Vp - (variance of the parents) - 395.376 Wr - (the mean covariance of the parents and the arrays) = 143.855 (the mean variance of the arrays) = 1747.664 (the variance of the means of the arrays) = 141.692 <: <1 H '1 I I 169 Table 20. Plant dry weight: array covariances, variances, and their differences (8x8, F1). Wr Vr Wr-Vr 257.5814 1803.9777 -1546.3963 259.5286 724.0860 —464.5574 558.5417 1153.4891 -594.9474 378.6945 936.3962 —557.7017 294.8829 613.4682 ~318.5853 251.9716 1669.7756 -1417.8041 -18.1539 439.6740 -457.8279 192.0286 716.9851 -524.9565 2175.0753 8057.8520 -5882.7766 Necessary statistic for plotting limiting parabola. Wr Vr 822.565 1803.978 521.134 724.086 657.751 1153.489 592.831 936.396 479.679 613.468 791.377 1669.776 406.088 439.674 518.573 716.985 3p - (variance of the parents) - 375.067 gr - (the mean covariance of the parents and the arrays) - 371.884 V5 - (the mean variance of the arrays) - 1007.231 Vr - (the variance of the means of the arrays) - 249.792 170 Table 21. Plant dry weight: array covariances, variances, and their differences (5x5, F2). Wr Vr Wr—Vr 13.9006 76.8356 -62.9350 8.7019 104.0901 -95.3883 259.7898 345.1188 -85.3290 250.0065 285.2190 -35.2124 -38.2374 26.8669 -65.1043 Total 494.1614 838.1304 -343.9690 Necessary statistics for plotting limiting parabola. Wr Vr 130.304 76.836 151.664 104.090 276.161 345.119 251.054 285.219 77.052 26.867 (variance of the parents) - 220.982 Wr - (the mean covariance of the parents and the arrays) - 98.832 (the mean variance of the arrays) - 167.626 (the variance of the means of the arrays) - 52.765 4 '0 I << '1'! ll 171 Ihbkgzz. Number of main stem branches: array covariances, variances, and their differences (5x5, F1). Wr Vr Wr-Vr 2.1284 4.0740 -1.9457 -.3316 ' 3.1058 -3.4373 .9712 1.0927 -.1216 Total 3.1091 9.0735 -5.9644 Necessary statistics for plotting limiting parabola. Wr Vr .859 .564 2.309 4.074 2.016 3.106 .557 .237 1.196 1.093 Yp (variance of the parents) 8 1.309 Hr - (the mean covariance of the parents and the arrays) = .622 Vr (the mean covariance of the arrays) = 1.815 Vr - (the variance of the means of the arrays) = .515 172 Tdflfi It Number of main stem nodes: array covariances, variances and their differences (5x5, F1). Wr Vr Wr-Vr .4411 .3640 .0771 1.1018 1.1899 -.0881 .9919 1.0924 -.1005 1.4052 1.2930 .1122 1.1143 .8216 .2927 Total 5.0542 4.7608 .2934 Necessary statistics for plotting limiting parabola Wr Vr .770 .364 1.393 1.190 1.334 1.092 1.452 1.293 1.157 .822 Vp - (variance of the parents) - 1.630 Er - (the mean covariance of the parents and the arrays) = 1.011 V: - (the mean variance of the arrays) 8 .952 Vr - (the variance of the means of the arrays) ' .669 173 Thbh324. Total number of nodes: array covariances, variances, and their differences (8x8, F1). Wr Vr Wr-Vr 422.8405 1891.4584 -1468.6179 338.8424 420.8217 -81.9793 306.7417 952.9258 -646.1841 257.5169 546.7503 -289.2333 334.5537 444.8759 -110.3222 401.3598 1781.1960 -1379.8362 -8.6904 253.3750 -262.0654 80.9154 343.3929 -262.4775 Total 2134.0800 6634.7960 -4500.7159 Necessary statistics for plotting limiting parabola. Wr Vr 803.436 1891.458 378.967 420.822 570.272 952.926 431.964 546.750 389.648 444.876 779.666 1781.196 294.059 253.375 342.333 343.393 Vp - (variance of the parents) - 341.276 fir - (the mean covariance of the parents and the arrays) = 266.760 75 - (the mean variance of the arrays) - 829.349 Vr - (the variance of the means of the arrays) = 347.593 Table 25. Total number of nodes: 174 array covariances, variances and their differences (5x5, F2). Wr Vr Wr-Vr 54.5366 154.3764 -99.8397 117.2463 125.1923 -7.9460 103.9059 177.3033 -73.3974 146.5915 230.8118 -84.2203 80.4161 83.6692 -3.2531 Total 502.6964 771.3530 -268.6565 Necessary statistics for plotting limiting parabola. Wr Vr 138.122 154.376 124.383 125.192 148.024 177.303 168.889 230.812 101.685 83.669 Vp - (variance of the parents) = 123.580 yr - (the mean covariance of the parents and the arrays) = 100.539 V5 - (the mean variance of the arrays) - 154.271 Vr - (the variance of the means of the arrays) = 82.889 175 lhbh326. Length of internode: array covariances, variances, and their differences (8x8, F1). Wr Vr Wr-Vr 681.5488 558.2203 123.3285 1387.0026 1992.3983 -605.3937 1272.0288 1606.1485 -334.1197 985.3777 988.3466 -2.9690 954.3713 1182.6354 -228.2640 1170.5062 1668.3578 -497.8516 391.4162 164.8066 -226.6096 636.3754 375.4472 -260.9282 Total 7478.6270 8536.3607 -1057.7337 Necessary statistics for plotting limiting parabola. Wr Vr 875.903 558.220 1654.784 1992.398 1485.751 1606.149 1165.488 988.347 1274.907 1182.635 1514.251 1668.358 475.927 164.807 718.336 375.447 2p - (variance of the parents) - 1374.379 gr - (the mean covariance of the parents and the arrays) = 934.828 V5 - (the mean variance of the arrays) - 1067.045 Vr - (the variance of the means of the arrays) - 676.769 176 Tdflg 2L, Length of internode: array covariances, their differences (5x5, F2). Wr Vr Wr-Vr 206.4770 104.2895 102.1875 684.4725 708.3345 —23.8620 518.2200 397.6907 120.5293 460.0102 599.8999 -139.8898 Total 2189.9007 2256.3851 -66.4844 Necessary statistics for plotting Wr Vr 280.427 104.290 580.030 446.170 730.835 708.335 547.612 397.691 672.573 599.900 (variance of the parents) - 754.050 (the mean covariance of the parents and the arrays) limiting parabola. (the mean variance of the arrays) - 451.277 (the variance of the means of the arrays) - 289.059 variances and 437.980 Table 28. Total <