PERFORMANCE OF DOUBLE CROSS HYBRIDS IN SOME EGYPTIAN COTTON GENOTYPES (Gossypium barbadense)

Five Egyptian cotton genotypes were used as parents and crossed by diallel mating design to produce ten single crosses and 15 double crosses across three succeed seasons (2016, 2017 and 2018) at the Agricultural Experiments and Research Station, Faculty of Agriculture, Cairo University, and Sakha Agricultural Research Station, Egypt. Out of five parents three were long staple, BBB (Big Black Boll), Australian and G97, while the other two were extra-long staple (Giza 92 and G96). Alpha lattice design (5 x 6) was used to evaluate 30 genotypes (five parents, ten single crosses and 15 double crosses). Studied traits were; seed and lint cotton yield, lint %, boll weight and No. of bolls/plant. Ordinary model of restricted maximum likelihood (REML) method was also preceded for all the genotypes. Estimate the efficiency of alpha lattice and restricted maximum likelihood models (REML) analysis in increasing the precision of the present trials The studied traits showed highly significant differences (p≤0.01) between genotypes. Moreove, mean square due to genotypes was partitioned into parents (P), single crosses (S), double crosses (D), P vs. S and P vs. D. ALL previous parts (i.e. P, S, D, P vs. S and P Vs. D) exhibited highly significant differences (p≤0.01) for most of studied traits. Concerning ANOVA table for double crosses showed that line general was either significant or highly significant for No. of bolls/plant, boll weight and seed cotton yield/kentar suggesting the presence of the additive variance in the inheritance of these traits subsequently selection would be efficient in improvement these traits. Also, 2-line arrangement was highly significant (p≤0.01) for all traits except for lint percentage (L%) suggesting the presence of the non-additive variance in the inheritance of these traits. 3-line arrangement was highly significant (p≤0.01) for all traits except for lint percentage (L %) indicating the contribution of the additive by dominance interaction including all three factors or higher order interaction except all dominance types. Positive and significant heterotic effects were detected for single and double crosses in most traits. With respect to lint cotton yield five single crosses showed positive and significant heterotic effect relatively to mid parent (Australian x BBB, Australian x G97, BBB x G97, BBB x G92 and BBB x G96) with amount of heterosis 18.18, 15.43, 18.64, 21.3 and 29.45%, respectively. With respect to seed cotton yield four crosses showed positive and significant mid parent heterotic (Australian x BBB) (G97 x G92), (Australian x G92)( BBB x G97), (Australian x G92)( BBB x G96) and (Australian x G92)( G97 x G96) with amount of heterosis 8.97, 10.26, 15.73 and 8.76%, respectively. The relative efficiency of alpha lattice design compared with a conventional RCBD was greater than one for all studied traits indicated that the alpha lattice results in a smaller error variance and it adjusts genotype means for block effects. In addition to that the relative efficiency is less than one for boll weight indicated that alpha lattice design is less efficient than the RCBD and means are not adjusted for block effects. Alpha lattice design was applied in the third season. The coefficient of variation (CV %) and standard error (SE) for the studied traits were calculated by ANOVA and ordinary Reml (Restricted maximum likelihood) between the genotypes and the highest CV% and standard error values were detected, followed by alpha lattice while the lowest value of these estimates were recorded by ordinary Reml indicated Reml has superior accuracy and precision compared to ANOVA for estimating variance components.