MATERIALS AND METHODS

The material for present investigation was developed at winter maize nursery, Amberpet, Hyderabad during 2004-2005.

Table 1:	Parental lines used in the present study

Fifteen white maize inbred lines including three local (WI-9, W-6 and W-7) and twelve exotic (GLET-7, GLET-27, CML-77, CML-79, CML-111, CML-138, CML-173, CML-213, CML-214, CML-240, CML-244 and CML-463) were crossed to three testers W₃, W₅ and W₃xW₅ in a linextester fashion. The tester genotypes were developed at maize Research Station, Pahalgam. The inbred testers W₃ and W₅ were elite inbred lines with good combining ability whereas, W₃xW₅ was a promising high yielding hybrid released by SKUAST-K. The source of germplasm lines used is given in Table 1. The parents (both lines and testers) and test crosses, (63 entries in all) were evaluated at two diverse locations in Kashmir valley viz., high altitude Rice Research Station, Larnoo and Regional Research Station, Wadura which represents distinct climatic regions. The design used was RBD with three replications at each location. Each genotypes was represented by two rows of 2 m length with inter and intra-row spacing of 65 and 20 cm, respectively. Data was recorded on six quantitative traits including grain yield plot-1, 100-seed weight, kernel rows ear^-1, ear length, ear diameter and plant height, from 10 randomly selected competitive plants from each replications. The data was analysed as per linextester procedure of Singh and Chowdary (1999) using WINDOSTAT software.

RESULTS

The pooled analysis of variance (Table 2) revealed significant mean squares due to progenies, lines, testers, crosses and parents v/s crosses for all traits except kernel rows ear^-1 for which mean squares due to lines was non-significant, indicating substantial variability in parental lines for these traits. Environmental component was significant for all traits but progeniesxenvironment interaction was non-significant for ear diameter and 100-seed weight. Comparatively mean squares due to testers was greater than lines for all traits except 100-seed weight. The analysis of variance for combining ability (Table 3) revealed significant variances due to GCA and SCA with greater magnitude of latter. Correspondingly the dominance variance was greater than additive component for all traits. The degree of dominance was in the range of over dominance for all traits.

The estimates of general combining ability (Table 4) revealed that the parents, CML-244, CML-79 and CML-214 were good general combiners for grain yield. CML-214 was a good general combiner for all traits. For 100-seed weight GLET-27 was the best general combiner followed by CML-210 and CMl-79. For kernel row ear^-1 CML-214, W-7 and CML-244 were good general combiner. For ear length CML-214, W-6 and CMl-111 were good general combiners for plant height, CML-463, CML-214 and GLET-27 were best general combiners in order of merit. Among testers, W₃ was a good general combiner for plant height and grain yield while as W₅ was a good general combiner for ear length, kernel rows ear^-1 and 100-seed weight and W₃xW₅ was a good combiner for ear diameter and grain yield.

The best crosses identified on the basis of per se performance and SCA effects are presented in Table 5.

Table 2:	Analysis of variance pooled over environments for six quantitative traits in maize
*,**, Significant at 5 and 1% level of significance, respectively

Table 3:	Pooled Analysis of variance for combining ability in a linextester analysis in maize
*,**, Significant at 5 and 1% level of significance, respectively

Table 4:	Pooled estimates of general combining ability effects of lines and testers in a linextester analysis in maize
*,**, Significant at 5 and 1% level of significance, respectively

Table 5:	Best crosses identified on the basis of per se performance and SCA effects in maize

For grain yield, CML-173xW₃ had highest mean performance followed by GLET-27xW₃ while as on the basis of SCA effects CML-213x(W₃xW₅), CML-79x(W₃xW₅) and CML-244xW₃ were the best crosses in order of merit. The parents CML-244, CML-240 and CML-214 were most frequently one of the parents in the best crosses besides possessing good GCA effects.

DISCUSSION

The success of plant breeding operation relies heavily on genetic variation. In fact plant breeding uses selection from the available genetic variability for crop improvement (Asins, 2002). The present study revealed substantial variability in parents as well as crosses which is encouraging for isolation of desirable genotypes using appropriate selection schemes. There was preponderance of SCA variances indicating greater role of non-additive components of variance in the inheritance of the traits studied. Similar, results in maize have been reported for yield and yield components traits by Dodiya and Joshi (2002) and Kumar et al. (2005). Thus hybrid breeding can be an efficient breeding strategy for improvement of grain yield and related traits in maize. In fact heterosis breeding has been the corner stone of maize improvement at national and international level. However, the stability of yield superiority needs to be tested over space and time. Recurrent selection can also be an efficient procedure for exploitation of both additive components of variance (Doerksen et al., 2003). Such a procedure increases the frequency of desirable alleles while maintaining the genetic variation without reaching a dead end. Thus both selfed progeny recurrent selection and reciprocal recurrent selection used individually or simultaneously in combination should permit benefits of both additive and non-additive gene effects in test cross and per se evaluation to be realised (Dhillon, 1991). The progress from simultaneous use of both procedures will be the total of expected gain from these procedures individually (Hallauer and Miranda, 1988). Popi and Kannenberg (2001) opined that under such schemes there is an increase in non-additive component relative to additive component. However, they were able to show that combining ability estimates are not changed substantially.

The average degree of dominance was in the range of over-dominance which further indicates the suitability of hybrid breeding. In fact, maize contains more over dominant loci as compared to other cereals even though all share common descent. The higher estimates for dominance variance also compensates for any upward bias due to linkage disequilibrium. Such overshadowing effect of non-additive variance has also bean reported by Das et al. (1997) and Suneetha et al. (2000). Furthermore, the present study revealed differential interaction of GCA and SCA variances. Whereas, both GCA and SCA variances interacted significantly for plant height, kernel rows ear-1 and 100-seed weight, only SCA interacted significantly with the environment and neither GCA nor SCA interacted with environment for ear diameter and grain yield indicating variance influence of environment on additive and non-additive components. Comparatively environmental interaction of non-additive component was greater than additive component for all traits, except 100-seed weight for which GCAxE was greater than SCAxE interaction. Similar results have been reported in maixe by Joshi et al. (1998), Zelleka (2000) and Betran et al. (2003). Such results further establish that need to carry out variance component studies across environments to get unbiased estimates.

A number of parental lines and testers were observed to posses good general combining ability effects for various traits especially CML-214, CMl-244 and CML-240, which not only had good GCA but also entered into specific cross combinations exhibiting superior mean performance and SCA effects for grain yield and related traits. These lines can be used in planned hybridzation programme to broaden the genetic base of local germplasm. Furthermore, the crosses recording superior performance can also serve as a source of improved lines with desirable alleles for a trait.