MATERIALS AND METHODS

Site Characteristics
This study was carried out during Spring and Summer of 2006 at the Miando'ab Agricultural Research Station, 46°09' N, 36°58' E and 1371 m altitude, in the province of West Azerbaijan, North West of Iran. An open pan evaporation class A was used to avoid of side effects of climatically parameters on the irrigation periods. The pan was installed in a 15 m distance apart from the experiment. Each period of irrigation in the non-stressed treatments was conducted after 90±5 mm evaporation of the pan, which was approximately equal to 10 days evaporation during June-July and for stressed plots in a 20 days intervals equal to 180±5 mm open pan evaporation during different growth phases.

Experimental Design
A split plot arrangement in randomized complete block design with four main plots, different levels of irrigation, including complete irrigation, stress at vegetative phase, stress at reproductive phase and stress at grain filling period and eight sub plot, consist of eight medium maturity hybrids of corn (SC 604, SC 647, SC 686, SC 687, SC 688, SC 689, SC 690 and SC 691), with three replications was performed.

Data Collected
Each hybrid was raised in two rows of 5 m length with a spacing of 60 cm between the rows and 20 cm between plants within the rows. Recommended package of practices were followed to raise the crop and normal cultural practices were followed through out the crop growth. Five plants were randomly selected from each hybrid, replication-wise for recording observations on fifteen characters, viz., days to silk emergence, plant height, ear length, number of rows per ear, number of kernels per row, number of branches per tassel, total number of leaves, number of leaves per ear, number of dry leaves, kernel depth, green cover percentage, leaf rolling, 300 kernels weight, single plant yield and Anthesis to Silking Interval (ASI).

Data Analysis
Standard statistical procedures were used for the analysis of variance and mean comparisons of the selected cultivars were conducted using Fisher’s LSD. The phenotypic correlation coefficients for yield and yield components were evaluated using the formula suggested by Jha et al. (1998). Further, the path coefficient analysis which measure the direct influence of one variable upon another and permits the separation of the simple correlation coefficient into components of direct and indirect effects was done according to Wright (1921) and Dewey and Lu (1959).

RESULTS AND DISCUSSION

The results of analysis of variance showed highly significant differences for most of the traits (Table 1). Effects of drought stress at different stages of growth, on traits such as 300 kernels weight, ASI, number of kernels per row, number of branches of tassel, green cover percentage and grain yield; and on characters like number of kernel rows, kernel depth, plant height and ear length, were significant at 5 and 1% levels, respectively. Therefore, the performance of aforementioned traits in destined hybrids under drought stress treatments (vegetative, reproductive and grain filling periods) toward normal condition has changed. There were significant differences between hybrids regarding to 300 kernels weight, ASI, number of kernel rows, number of branches of tassel, number of kernel rows, green cover percentage and grain yield at 1% significant level and also for number of days to silk emergence, ear length and number of top leaves at 5% level, showing presence of variation for these traits among hybrids.

Table 1:	Phenotypic variation in different characters of maize hybrids
ns: Not significant; *: Significant at p = 0.05; **: Significant at p = 0.01

Table 2:	Performance of each character of maize hybrids in normal situation and different levels of drought stress

Number of days to silk emergence and ASI had lowest and highest coefficients of variation, respectively. In general, results of ANOVA showed that there were a dramatically genetic variation among investigated hybrids and it could be exploited to select genotypes for drought prone environments.

Mean grain yield among the trials ranges from 9.9 to 12.3 t ha^-1 (Table 1, 2). Interaction of hybridxstress levels was significant for ASI and ear length (data not shown). Overall mean of different characters for each single cross have showed in Table 3. Hybrid number six (SC 686) produced higher grain yield and lowest ASI (2.25 days). This genotype was superior over others regarding to some traits such as 300 kernel weight, number of kernel per row, number of rows per ear, kernel depth, green cover percentage and ear length. Hybrid number five (SC 687), showed highest value of ASI (3.42 days) and lowest number of kernels per row (41.58). Westgate (1994) and Edmeades et al. (1994) reported negative correlation between ASI and number of kernels per row. However, with ASI duration more than 8 days, number of kernels per row will be decrease significantly to produce even seedless ears.

Simple correlation coefficients among traits at different moisture contents were shown in Table 4-7. The coefficients of correlation for grain yield vs. 300 kernels weight, green cover percentage, number of kernel per row, leaf number and plant height were 0.63, 0.77, 0.57, 0.54 and 0.42, respectively and significant. Whereas correlation for grain yield vs. ASI and number of branches of tassel were negative and significant (Table 4). Bolanos and Edmeades (1996) found positive and significant coefficients of correlations between grain yield vs. number of kernel per row and one-hundred kernels weight at drought stress condition.

In the present study, correlation between grain yield vs. 300 kernels weight (0.51), number of kernel per row (0.72), kernel depth (0.54) and green cover percentage (0.41) were positive and significant at vegetative phase stress (Table 5). Also, correlation coefficient between grain yield and ASI was negative (-0.53) and significant (p = 0.05). Abrecht and Carberry (1993) in their study on effect of drought stress on different properties of maize reported that, moisture stress at vegetative phase increases association between grain yield and ASI.

Correlation coefficients between grain yield against 300 kernels weight (0.63), number of kernels per row (0.66), plant height (0.63) and ear length (0.41) were significant. Correlation between grain yield and ASI (-0.84), as well as grain yield and number of branches of tassel (-0.81) were significant (p = 0.01).

Table 3:	Response of maize hybrids

Table 4:	Phenotypic correlation coefficients between grain yield and some yield components and growth characters of maize hybrids under drought stress at normal condition
*Significant at p = 0.05; **significant at p = 0.01

Table 5:	Phenotypic correlation coefficients between grain yield and some yield components and growth characters of maize hybrids under drought stress at vegetative phase
*Significant at p = 0.05; **significant at p = 0.01

Table 6:	Phenotypic correlation coefficients between grain yield and some yield components and growth characters of maize hybrids under drought stress at reproductive phase
*Significant at p = 0.05; **Significant at p = 0.01

Table 7:	Phenotypic correlation coefficients between grain yield and some yield components and growth characters of maize hybrids under drought stress at grain filling phase
*Significant at p = 0.05. **Significant at p = 0.01

Meanwhile, correlation of ASI with yield components and green cover percentage was negative and significant, but with number of branches of tassel and days to silk emergence was positive and significant (Table 6). As shown in Table 7, stress at kernel filling period brought positive and significant correlation between grain yield and 300 kernel weight (0.78), number of kernel per row (0.67), kernel depth (0.42) and green cover percentage (0.53). According to Claassen and Shaw (1970), stress at kernel filling period has not any effect on kernel number, but due to low leaf photosynthesis and less remobilization of assimilate toward seeds at this condition, weight of kernels will be decreased. Three key functions for water deficiency in maize are intensity and duration of drought stress, rate of stress development and phonological timing of stress (Galeta et al., 2005; Bolanos et al., 1993). Present analysis and several investigations show that, anthesis and silking phase (two weeks before emergence of silks), are most tender stages to drought and stress in these scenes will induce prolonged interval between anthesis to silking, decline of kernel number and thereby, reduced grain yield (Abdmishani et al., 2004; Cavalieri and Smith, 1985; Downey, 1977; Kabdal et al., 2003; Venugopal et al., 2003; CIMMYT, 2000).

The simple correlation coefficients between grain yield and some yield component and growth characters of under study single cross hybrids of maize were partitioned into direct and indirect effects using path analysis (Table 8-11). The estimated residual effects had indicated that the characters chosen for path analysis were appropriate. Moreover, the values of direct effects indicating the minimum inflation due to multi co-linearity (Mohammadi et al., 2003). The path analysis on grain yield of drought stress at reproductive phase (Table 8), revealed that 300 kernels weight (0.33) recorded high positive direct effect on grain yield and its indirect effects on grain yield via kernels/row (0.32) and ASI (0.85) was positive and via days to silk emergence was negative (-0.86).

Table 8:	Phenotypic correlations with grain yield, direct and indirect effects of yield components at reproductive phase stress conditions
Bold values have direct effects. *Significant at p = 0.05. **Significant at p = 0.01, Residual effect: 0.107, R2 = 0.93

Table 9:	Phenotypic correlations with grain yield, direct and indirect effects of yield components at grain filling phase stress conditions
Bold values have direct effects. *Significant at p = 0.05. **Significant at p = 0.01, Residual effect: 0.107, R2 = 0.81

Table 10:	Phenotypic correlations with grain yield, direct and indirect effects of yield components at vegetative phase stress conditions
Bold values have direct effects. *Significant at p = 0.05. **Significant at p = 0.01, Residual effect: 0.004, R2 = 0.72

Table 11:	Phenotypic correlations with grain yield, direct and indirect effects of yield components at non-stress conditions
Bold values have direct effects. *Significant at p = 0.05. **Significant at p = 0.01, Residual effect: 0.102, R2 = 0.94

The ASI had negative direct effect (-1.15) and negative indirect effects on grain yield via 300 kernel weight (-0.24) and kernels/row (-0.30). Direct effect of this trait proved to have impact on its correlation with grain yield at mentioned conditions. Therefore, simultaneously screening for ASI and days to silk emergence might be accomplished to select shorter ASI and drought tolerant single crosses of maize.

Considering drought stress at grain filling stage, 300 kernels weight (0.69), kernels/row (0.45), green cover percentage (0.17) and leaf rolling (0.23) had direct effect on grain yield (Table 10). On the other hand, indirect effects of green cover percentage via 300 kernel weight and kernels per row was more than its direct effect on grain yield. So, it could be concluded that, at grain filling period of drought prone condition, high green cover percentage plants with broad blade level are capable to produce more photosynthetic assimilates and remobilization them from vegetative organs to the grains and ears, which ultimately increase 300 kernels weight. The obtained outcomes are in consistent with those reported by Abdmishani et al. (2004).

At vegetative period drought stress, 300 kernels weight recorded medium level of direct effect and low level of indirect effect via kernels per row on grain yield. The path analysis revealed that, number of kernels per row (0.24) recorded positive direct effect and positive indirect effects via 300 kernels weight, kernel depth and green cover percentage on grain yield. Regarding to positive direct and indirect effects of this trait on grain yield and its components, it could be exploited for selection at drought resistance programs. The results presented by Singh and Singh (1993) emphasized that number of kernels per row; kernels weight and number pf ear per plant are the most important traits, respectively.

Number of kernels per row proved to have a high direct effect on grain yield (0.82) and its indirect effects via 300 kernels weight, number of kernels per row and kernel depth exhibited positive and noteworthy.

From the foregoing discussion, it may be concluded that, drought resistance dedicating traits are polygenic since drought tolerant single crosses had got highest yields at all moisture stress conditions (SC 686, SC 690 and SC 604, respectively). Moreover, the present study revealed that phenotypic correlations of important traits with each other and with grain yield at normal and stressed conditions mostly were concurrent. Though, values of correlations of ASI, number of kernels per row, days to silk emergence and ear length with grain yield at drought prone conditions varied considerable, which showing divers response of hybrids at different situations of moisture, since, sign and value of association of these traits affected sorely by media and stage of growth as well as intensity of drought stress.