Eman Salah El-Din Abd El-Aziz, <br/><br/>
Untraditional Statistical Methods for Evaluating Diallel Cross   Outputs Towards Some Desirable Egyptian Cotton Traits/ 
 الطرق االاحصائية غير التقليدية لتقييم نواتج الهجن التبادلية فى اتجاه بعض الصفات المرغوبة فى القطن المصرى/ 
Eman Salah El-Din Abd El-Aziz ; Supervisors: Dr. Deyaa Ahmed El-Kadi, Dr. Mohamed Abd El-Maboud Abd El-Shafi, Dr. Talaat Ahmed El-Feki. 
 - 129 pages :  illustrations ;   25 cm. +   CD. 
<br/><br/>Thesis (Ph.D)-Cairo University, 2024. 
<br/><br/>Bibliography: pages 118-129. 
<br/><br/> Five Egyptian cotton genotypes (Australian, BBB, G97, G92 <br/>and G96) were used as parents in this study. Three of these <br/>varieties were long staple, BBB (big black boll) (P1), Australian <br/>(P2) and G97 (P3) while the other two varieties were extra-long <br/>staple, Giza 92 (P4) and G96 (P5).<br/>Diallel mating design was used to produce 10 single crosses <br/>and 15 double crosses across two seasons (2016 and 2017) at the <br/>Agricultural Experiments and Research Station, Faculty of <br/>Agriculture, Cairo University, Egypt. The final experiment was <br/>conducted in the third growing season (2018) using alpha lattice <br/>design (5 x 6) to evaluate 30 genotypes (parents, single crosses and <br/>double crosses) at Sakha Agricultural Research Station, Agriculture <br/>Research Center, Egypt.<br/>The present study was therefore, designed to: <br/>(1) Generate new forms of cotton genotypes having high seed <br/>cotton yield potential and one or more desired of traits. <br/>(2) Characterize the desirable traits and existing cotton parents and <br/>their crosses for early maturity and determine the <br/>interrelationship between early maturing, yield and fiber <br/>quality traits. <br/>(3) Determine various genetic parameters for yield, its components <br/>and quality traits in crosses of Egyptian cotton. <br/>(4) Estimate the efficiency of alpha lattice and restricted maximum <br/>likelihood models (REML) analysis in increasing the precision <br/>of the present trials. <br/>110<br/>First Season<br/>In first season 2016, the five parents were planted and mated <br/>in a diallel mating design excluding reciprocals to obtain 10 single <br/>crosses. Seeds from each parental genotype were sown in two rows <br/>for each parent. The row was 7 meters long and 60 cm apart and <br/>distance of 70cm between the plants on hills. Later the plants were <br/>thinned to one. Emasculation and crossing were made between the <br/>parents at flowering stage.<br/>Second Season<br/>In the second season 2017, F1 single crosses were grown to <br/>mate in a diallel mating design to produce double cross seeds with <br/>the restriction of no parent appeared twice in the same double cross <br/>combination to obtain 15 double crosses.<br/>Third Season<br/>In the final experiment (third season 2018), alpha lattice <br/>design (5x6) with three replications was used in evaluating 30 <br/>genotypes (the 5 parental genotypes, 10 F1`s single crosses and 15 <br/>double crosses). Each plot consisted of two rows. The rows were 4 <br/>meters long and 65 cm apart. Hills were spaced at 20 cm within <br/>rows and seedlings were thinned at two plants/hill. All cultural <br/>practices were followed throughout the growing season as usually <br/>done with ordinary recommendations for cotton culture.<br/>111<br/>The Studied Traits<br/>(1) Seed cotton yield/kentar (SCY/K), (2) Lint cotton <br/>yield/kentar(LY/K), (3) Boll weight (g) (BW), (4)Number of open <br/>bolls/plant (NO.B/P), (5) Lint percentage (%) (L%).<br/>Earliness traits<br/>(1) First fruiting node of first sympodium (F.F.N), (2) Days to first <br/>flower appearance (D.F.F), (3) Duration of the boll maturation <br/>(day).<br/>Fiber properties<br/>(1)Upper half mean (U.H.M), (2) Fiber strength (F.S), (3) <br/>Micronaire value (Mic). <br/>Statistical Analyses<br/>Statistical procedures for alpha lattice design were <br/>conducted according Cochran and Cox (1957). Ordinary Reml <br/>model analysis was described by Hough et al. 2010 and Piepho et <br/>al., 2012. Data were analyzed using Genstat (2011).<br/>The procedures of single crosses analysis were described by <br/>Griffing (1956) based on method 2, model I (fixed model) as <br/>outlined by Singh and Chaudhary (1985). The procedures of double <br/>crosses analysis (quadriallel) were carried out according to the <br/>procedure outlined by Singh and Chaudhary (1985). The GCA <br/>effects of parents and SCA effects of F1 crosses were calculated <br/>according to the method described by Griffing (1956) based on <br/>method 2, model I (fixed model) as outlined by Singh & <br/>112<br/>Chaudhary (1985). Estimates of heritability were determined <br/>according to the equations of Mather & Jinks (1982). <br/>The results could be summarized as follows:<br/>1. Earliness Traits<br/>1. Line general was highly significant (P≤ 0.01) for all traits <br/>suggesting the presence of the additive variance in the <br/>inheritance of these traits subsequently selection would <br/>be efficient in improvement these traits. Also, 2- line <br/>arrangement was highly significant (P≤ 0.01) for all <br/>traits, excluding days to opening 1st boll which was <br/>significant only (P≤ 0.05) suggesting the presence of the <br/>non-additive variance in the inheritance of these traits. <br/>Also, 3- Line arrangement was highly significant (P≤ <br/>0.01) for all traits excluding for days to opening first boll <br/>which was significant only (P≤ 0.05) indicating the <br/>contribution of the additive by dominance interaction <br/>including all three factors or higher order interaction <br/>except all dominance types.<br/>2. Highly significant differences for mean squares of P, C, <br/>D, P vs. C and P vs. D (P≤0.01) for all earliness traits. P <br/>vs. S was insignificant for position of first fruiting node <br/>and days to opening first boll and P vs. D was <br/>insignificant for position of first fruiting. Highly <br/>significant differences for mean squares of P, C, D, P vs. <br/>C and P vs. D (P≤0.01) for all earliness traits. <br/>113<br/>3. Double cross [(BBBxG97)(G92xG96)] was the earliest<br/>for position of first fruiting node and days to first flower <br/>and showed negative and significant mid parent heterosis <br/>with significant different from [(AustralianxG96) <br/>(BBBxG97)].<br/>4. Alpha lattice design was more efficient than conventional <br/>RCBD for most of studied traits and means are adjusted <br/>for block effects. Lowest values for CV and SE for traits <br/>were calculated by ordinary REML reflecting the <br/>accuracy and precision of this method compared to <br/>ANOVA for estimating variance components<br/>5. GCA mean squares were significant for both F1 and DC <br/>in all earliness traits except for days to first flower. SCA <br/>mean squares were highly significant only for F1 in two <br/>traits first fruiting node and days to first flower. <br/>GCA/SCA ratio revealed that GCA was higher than SCA <br/>in F1 for duration of the boll maturation while in DC, it <br/>was higher than SCA for first fruiting node and days to <br/>opening 1st boll.<br/>6. 4-line interactions with respect to particular arrangement <br/>of the parents in double crosses, (P1xP3) (P4xP5) was <br/>the best combination for all earliness traits.<br/>2. Yield and Yield Traits<br/>1. Genotypes included in the study had highly significant <br/>variation (p≤0.01) for all studied traits. genotypes was <br/>partitioned into parents (P), single crosses (S), double <br/>114<br/>crosses (D), P vs. S and P vs. D. ALL previous parts (i.e. P, <br/>S, D, P vs. Sand P Vs. D) exhibited highly significant <br/>differences (p≤0.01) for all studied traits, except lint <br/>percentage (L %) due to P and D. Also, P vs. S was <br/>insignificant for seed and lint cotton yield/kentar and P vs. <br/>D was insignificant for lint cotton yield/kentar.<br/>2. Line general was highly significant (p≤0.01) only for No. <br/>of bolls/plant and boll weight suggesting the presence of the <br/>additive variance in the inheritance of these traits <br/>subsequently selection would be efficient in improvement <br/>these traits, whereas it was significant (p≤0.05) only for seed <br/>cotton yield/kentar. Meanwhile, line general was <br/>insignificant for lint cotton yield/kentar. 2-line arrangement <br/>was highly significant (p≤0.01) for all traits except for lint <br/>percentage (L%) suggesting the presence of the non-additive <br/>variance in the inheritance of these traits. 3-line arrangement <br/>was highly significant (p≤0.01) for all traits except for lint <br/>percentage (L %) indicating the contribution of the additive <br/>by dominance interaction including all three factors or <br/>higher order interaction except all dominance types. <br/>3. The double cross (AustralianxG92) (G97xG96) had the <br/>highest mean performance value and showed positive and <br/>significant mid parent heterosis for most of yield traits.<br/>4. Alpha lattice design was more efficient than conventional <br/>RCBD for most of studied traits except for boll weight.<br/>Lowest values for CV and SE for traits were calculated by <br/>115<br/>ordinary REML reflecting the accuracy and precision of this <br/>method compared to ANOVA for estimating variance <br/>components. Also results showed great differences between <br/>standard error for different traits under RCBD and standard <br/>error (SE) of the same traits under alpha lattice design. The <br/>smaller values of SE difference for alpha lattice design helps <br/>to detect smaller differences for the comparisons of mean. <br/>5. GCA mean squares were highly significant (P≤0.01) for <br/>F1 and DC in all yield traits except for Lint% which was <br/>insignificant with DC. Specific combining ability (SCA) <br/>mean squares were highly significant (P≤0.01) for F1 and <br/>DC in all yield traits except for Lint% with F1 and No. of <br/>bolls/plant with DC. GCA/SCA ratio revealed that GCA <br/>was higher than SCA in F1 for seed cotton yield, lint cotton <br/>yield, Lint % and boll weight while in DC; it was higher <br/>than SCA for seed cotton yield and No. of bolls/plant<br/>6. 4-line interaction with respect to particular arrangements <br/>of the parents in double crosses, (P1xP2) (P4xP5) and <br/>(P1xP3) (P4xP5) formed the best combination for most <br/>yield traits.<br/>3. Fiber Quality Traits<br/>1. Statistical analysis revealed that the genotypes <br/>included in the study had highly significant variation <br/>(p≤0.01) for all studied traits except for upper half <br/>mean which showed insignificance.<br/>116<br/>2. 1-line general, 2-line arrangement and 3- line <br/>arrangements, all the parts exhibited highly <br/>significant difference except upper half mean.<br/>3. Mean squares due to genotypes was partitioned into <br/>parents (P), single crosses (S), double crosses (D), P <br/>vs. S and P vs. D. Previous parts (i.e. P, S, D, P vs. S <br/>and P vs. D) exhibited highly significant differences <br/>(p≤0.01) for all studied traits. P vs. S was <br/>insignificant for fiber strength.<br/>4. Double cross [(AustralianxG96)(G97xG92)] had the <br/>highest mean performance for most traits and showed <br/>positive and significant mid parent heterosis<br/>5. Alpha lattice design was more efficient than <br/>conventional RCBD for most of studied traits and <br/>means are adjusted for block effects. Coefficients of <br/>variation (CV%) of alpha lattice design were low as <br/>compared to RCBD for fiber length and fiber <br/>strength. Lowest values of the coefficients of <br/>variation (CV%) for all traits and lowest standard <br/>error were recorded for reml analysis in fiber length <br/>and fiber fineness which indicates high experimental <br/>precision.<br/>6. GCA mean squares were significant for fiber length <br/>and fiber strength in F1, while GCA mean squares <br/>were significant for fiber strength and fiber fineness <br/>(Micronaire reading) in DC. Specific combining <br/>117<br/>ability mean squares were only significant for F1 in <br/>both traits fiber strength and fiber fineness <br/>(Micronaire reading). GCA/SCA ratio revealed that<br/>GCA was higher than SCA in F1 for fiber length <br/>while in DC, it was higher than SCA for fiber length <br/>and fiber fineness (Micronaire reading).<br/>7. For 4-line interaction with respect to particular <br/>arrangements of the parents in double crosses, the <br/>best combination was (AustralianxG97) (G92xG96).<br/><br/>  تم استخدام خمسة اصناف من القطن المصري كأباء وهى BBB، أسترالي ، 97G، 92G ، 96G. أجريت التجارب في محطة تجارب كلية الزراعة في الموسمين األاول والثانى اما التجربة النهائية تم اجرائها في محطة تجارب سخا وذلك لتقييم 30 تركيب وراثى( 5أباء , 10 هجن فردية , 15هجين زوجى) باستخدام التصميم الشبكى 6x5 .الصفات التي تم دراستها هى محصول القطن الزهر والشعر, نسبة تصافي الحليج, متوسط وزن اللوزة,عدد اللوز المتفتح علي النبات ,نعومة التيلة, متانة التيلة  ,طول التيلة, موقع أول عقدة ثمرية , ميعاد ظهور أول زهرة ,ميعاد تفتح أول لوزة. تم عمل التهجين نصف الدائرى للحصول على 10 هجين فردى و 15 هجين زوجى.تم اجراء تحليل Analysis REML لتقييم كل التراكيب الوراثية . تم تقدير الكفاءة النسبية لتصميم design lattice alpha وكذلك Analysis REML في زيادة الدقة التجريبية مقارنة بتصميم RCBD 
<br/><br/>
<br/><br/><br/>Text in English and abstract in Arabic & English. 
<br/><br/><label>Subjects--Topical Terms: </label><br/>Cotton
<br/><br/><label>Subjects--Index Terms: </label>Gossypium barbadense L.  Single Crosses  Double Crosses  Alpha  Lattice Design  Ordinary Reml  Mid-Parent Heterosis  Combing  ability  Heritability  Order effect 
<br/><br/><label>Dewey Class. No.: </label>633.51