The combining ability estimates for yield , yield components and fiber traits in cotton were the ultimate aim of this investigation . In this study six cotton varieties i.e.: G. 45 , G.89 x Pima S6 , C.B.58 , TNB , Aust. 12 and G. 89 belong to Species Gossypium barbadense L. and their 45 double crosses were used for this purpose . In the 2008 growing season , these all genotypes ( six varieties and 45 double crosses ) were evaluated in a field trial experimental at Sakha Agriculture Research Station for following traits : seed cotton yield per plant (S. C.Y./P.), lint cotton yield per plant (L.C.Y./P.) , number of seeds per boll ( No. S./B.) , number of bolls per plant ( No. B./P.) , fiber strength ( F.S.) , fiber finenesses ( F.F.) , 2.5% span length ( 2.5% S.L.) , 50% span length ( 50% S.L.) and uniformity ratio ( U.R.%.) .
The results showed that the mean squares of genotypes were significant and highly significant for all the studied traits . Further partition of crosses mean squares to its component showed that the mean squares due to 1- line general , 2- line specific effect , 2- line arrangement , 3- line arrangement and 4- line arrangement were either significant or highly significant for most studied traits suggesting the presence of the additive and non- additive variance in the inheritance of these traits .
For the mean performances , the results showed that the cross [ ( P3 x P6) x ( P4 x P5 )] cleared the highest mean for seed cotton yield per plant , lint cotton yield per plant , the cross [ ( P3 x P4 ) x ( P5 x P6 )] showed the highest mean for number of seeds per boll . Also, the crosses [ ( P2 x P6) x ( P3 x P4)] , [ ( P2 x P3) x ( P4 x P6)] and [ (P2 x P3 ) x ( P4 x P5) ] were the best crosses for number of bolls per plant , fiber strength and fiber finenesses . In addition , the cross [ ( P1 x P6 ) x ( P2 x P4 ) ] was the superior for 2.5% span length and 50% span length , as well as , the cross [ ( P2 x P6 ) x ( P4 x P5) ] was the best cross for uniformity ratio .
The variety C.B.58 ( P3 ) was the best combiner for most of yield and yield components traits . . In the same time , the variety TNB ( P4 ) was the best and good combiner for most of fiber traits . The variety G.89 x Pima S6 ( P2) was the best combiner for ( No. B./ P.) while the variety Aust. 12 ( P5) was the good combiner for (U.R.% ) .
The results also cleared that , the two- line interaction effect ( S212) and ( S226) showed positive ( desirable ) effect for most of yield studied traits . Also ,( S234) was the best combination for ( F.S.) , ( 2.5% S.L.) and ( 50% S.L.) .Regarding the three – line interaction effect , the results showed that the combination ( S3256) was the best combination for ( S.C.Y./ P.) , ( L.C.Y./ P.) and ( No. B./P.) . In addition , ( S3345) was the best positive desirable effect for most of fiber traits. . Also , the combinations ( S3126) , ( S3346) and ( S3146) were the best combinations for ( No. S./B.) , ( F.S.) and ( F.F.) , respectively .. Moreover , ( S41256) was the best combination for ( 2.5% S.L.) , ( 50% S.L.) and ( U.R.%.) , as well as , ( S42345) and ( S42346) were the best combination for ( No. S./B.) , ( F.S.) and ( F.F.) .
The specific combining ability effects ( t2 (ij)(..) showed that the combination t2(36) (..) was the best for ( S.C.Y./P.) and ( L.C.Y./P.) , t2(15) (..) and t2 (12) (..) were the best for ( No. B./P.) and ( F.S.) , respectively . Also , the combination t2(16) (..) was the best for ( 2.5% S.L.) , ( 50% S.L.) and ( U.R,%) .
The specific combining ability effect t2(i.) (j.) cleared that the combination t2 (1.) (2.) , t2 ( 3.) (6.) and t2 (3.) (4.) were the best combinations for ( No. S./B.) , ( F.F.) and ( U.R.%) , respectively , as well as the combination t2 (2.) (6.) was the best effect for (2.5% S.L.) and ( 50% S.L.) .
The results revealed that the dominance genetic variance (σ2D) was positive and larger than those of additive genetic variance (σ2A ) for most of the studied traits . Concerning epistatic variance , additive by additive (s2AA), additive by dominance (s2AD), dominance by dominance (s2DD) and additive by additive by additive (s2AAA) were positive with considerable magnitude for most of studied traits. Therefore, it could be recommended the recurrent selection breeding program for improvement these traits with respect to these genetic materials.