The Influence of Water Stress on Yield and Related Characteristics in Inbred Quality Protein Maize Lines and Their Hybrid Progeny

(MPOM), (MLOM), (PWOM), Awassa (AWOM), Jimma (JMOM) and Kiboko (KBOM) research stations comprised optimum management (optimal fertilization and supplemental irrigation as needed to avoid water stress). Fertilizer rates at each location were adjusted to reflect the agronomic recommendations for each location. The trials were conducted during the summer (main cropping) seasons of the respective countries. Two experiments were grown under water stress during the winter (dry) seasons at Chiredzi, Zimbabwe (CHDS) and Kiboko, Kenya (KBDS) research stations. Plants experience water stress either when the water supply to their roots becomes limiting, or when the transpiration rate becomes intense. Water stress is primarily caused by a water deficit, such as a drought or high soil salinity. Each year, water stress on arable plants in different parts of the world disrupts agriculture and food supply with the final consequence: famine. Hence, the ability to withstand such stress is of immense economic importance. Plants try to adapt to the stress conditions with an array of biochemical and physiological interventions. This multi-authored edited compilation puts forth an all-inclusive picture on the mechanism and adaptation aspects of water stress. The prime objective of the book is to deliver a thoughtful mixture of viewpoints which will be useful to workers in all areas of plant sciences. We trust that the material covered in this book will be valuable in building strategies to counter water stress in plants.

. Locations and environments used to evaluate F 1 hybrids, with their characteristics and codes

Germplasm
Fifteen inbred lines were selected based on diverse pedigree backgrounds. These lines showed better combining ability in top-cross evaluations and per se performance across a range of tropical and subtropical environments (data not shown). Most of the lines are resistant/tolerant to major foliar diseases of the tropics (CIMMYT, 2004). Diallel crosses were made among the 15 inbred lines in the winter of 2006 at Muzarabani, Zimbabwe. Seeds from reciprocal crosses were bulked to form a set of 105 F 1 hybrids. The F 1 hybrids were evaluated along with two QPM (SC527Q and CML144/CML159//CML176) and one normal maize (SC633) hybrid checks in all experiments conducted in Kenya, Zambia and Zimbabwe, and two normal maize (BH540 and BH541) and one QPM (BHQP542) hybrid checks in all experiments conducted in Ethiopia.

Experimental design and field measurements
All experiments were laid out as 9 x 12 alpha-lattice designs (Patterson and Williams, 1976) with two replications (Table 1). Measurements were recorded on well-bordered plants by excluding the plant nearest to the alley of each row. Days to anthesis and silking w e r e c a l c u l a t e d a s t h e n u m b e r o f d a y s f r o m p l a n t i n g t o 5 0 % p o l l e n s h e d a n d s i l k emergence. Anthesis silking interval was calculated as the difference between days to silking and anthesis (ASI = DS -DA). Two weeks after pollen shed, plant height and ear height were measured as the distance from ground level to the first tassel branch or to the node bearing the main ear. Number of ears per plant was obtained by dividing the number of ears by number of plants harvested. An ear was counted if it had at least one www.intechopen.com Water Stress 202 fully developed grain. Grain weight from all the ears of each experimental unit was measured and used to calculate grain yield (expressed in ton ha -1 and adjusted to 12.5% moisture content).

Statistical analysis
Before data analysis, anthesis-silking interval (ASI) was normalized using ln (1 0 ) ASI  as suggested by Bolanos and Edmeades (1996). Analysis of variance per environment was conducted with the PROC MIXED procedure in SAS (SAS, 2003) considering genotypes as fixed effects and replications and blocks within replications as random. Entry means adjusted for block effects generated from individual location analyses according to a lattice design (Cochran and Cox, 1960) were used to perform across environments combined analyses using PROC GLM in SAS (SAS, 2003) and combining ability analysis using a modification of the DIALLEL-SAS program (Zhang and Kang, 1997). GCA effects of the parents and SCA effects of the crosses were estimated following Griffing's Method IV (crosses only) and Model I (fixed) of diallel analysis (Griffing, 1956). Combined analyses of variance were conducted for each trait that showed significant entry mean squares in individual environment analysis. Combining ability was analyzed, and GCA and SCA effects were estimated accordingly. The mean squares for hybrids and environments were tested against the mean squares for hybrid x environment (E) as error term while hybrid x E interactions mean squares were tested against pooled error. Since means (over replication) of each of the genotypes were used for combined analysis of variance, estimate of pooled error mean squares were calculated following the procedure of Dabholkar (1999)  error mean square at i th environment, respectively, n is the number of environments and r is the number of replications in each environment. The significance of GCA and SCA sources of variation was determined using the corresponding interactions with the environment as error terms. Error mean squares calculated above were used to test the significance of GCA and SCA interactions with environment; because the combining ability mean squares were calculated based on entry means of each genotype from each environment (Griffing, 1956;Singh, 1973;Dabholkar, 1992). For GCA effects of the inbred lines, the restriction 0 gi   was imposed. Significance of GCA effects was determined by the t-test, using standard errors of GCA effects (Griffing, 1956;Singh and Chaudhary, 1985).

Results
Analysis of variance for each environment revealed the existence of significant differences among hybrids for most traits except anthesis-silking interval at Harare, Mpangwe and Pawe optimal (Table 2). Mean squares due to GCA were highly significant for all traits studied at all environments. SCA effects were also significant for most traits. Mean grain yields for the QPM hybrids (excluding the checks) ranged from 0.6 t ha -1 under severe water stress at Chiredze to 8.4 t ha -1 under optimum management at Mpongwe (Table 3). At Kiboko, average grain yield of the hybrids tested under water stress was 35.7% of grain yield under optimal conditions (KBOM).

www.intechopen.com
The Influence of Water Stress on Yield and Related Characteristics in Inbred Quality Protein Maize Lines and Their Hybrid Progeny 203 HAOM=Harare optimal, RAOM=Rattray optimal, MPOM=Mpongwe optimal, BKOM=Bako optimal, MLOM=Melkasa optimal, PWOM=Pawe optimal, AWOM=Awassa optimal, JMOM=Jimma optimal, CHDS=Chiredzi stress, KBDS=Kiboko stress, KBOM=Kiboko optimal* P≤ 0.05 ; ** P≤ 0.01; DF= degrees of freedom; GY= grain yield; AD= days to anthesis; DS= days to silking; ASI= anthesis-silking interval; PH= plant height; EH= ear height; EPP= ears per plant Table 2. Mean squares for hybrids, general (GCA) and specific (SCA) combining ability for grain yield and agronomic traits in stressed and optimal environments, 2006 -2008 Combined analysis of variance across water stress environments revealed highly significant mean squares due to environments and hybrids for all traits analyzed (Table 4). Mean grain yield across water stress environments ranged from 0.3 to 3.7 t ha -1 with a mean of 1.8 t ha -1 . Higher grain yields were recorded for VL052 x VL05561 (3.7 t ha -1 ), VL05561 x CML159 (3.5 t ha -1 ), VL054178 x VL06375 (3.4 t ha -1 ), VL05482 x VL05561 (3.3 t ha -1 ) and VL054178 x VL05561 (3.0 t ha -1 ). Mean grain yield across water stress environments (Table 4) was 27.4% of the mean grain yield across optimal environments (Table 5). Mean days to anthesis was 92.3 with a range of 82.8 -103.5. Days to silking ranged from 83.7 to 120.0 d with a mean of 102.0. Anthesis-silking interval ranged from 0.4 to 21.4 with a mean of 9.7. Ears per plant ranged from 0.10 to 0.88 with a mean 0.50. Combining ability analysis revealed nonsignificant GCA mean squares for grain yield but significant GCA mean squares for days to anthesis and silking, anthesis-silking interval and ears per plant. SCA mean squares, however, were not significant for all traits. Hybrid x E, GCA x E and SCA x E interaction mean squares were significant for all traits tested. Across optimal environments, the effects of environments, hybrids, GCA and SCA were highly significant for all the traits evaluated (Table 5). Grain yields ranged from 1.8 to 9.4 t ha -1 with a mean of 6.5 t ha -1 . The highest yielding hybrids were VL05483 x CML491 (9.4 t ha -1 ), CML511 x CML491 (8.8 t ha -1 ), VL05561 x CML491 (8.7 t ha -1 ), CML159 x CML491 (8.5 t ha -1 ) and VL054178 x CML491 (8.1 t ha -1 ). Mean days to anthesis was 73.8 with a range of 66.9 -80.4. Days to silking ranged from 68.9 to 82.8 with a mean of 75.1. Mean plant and ear height was 225.5 and 110.9 cm with ranges of 189.0 -248.4 cm and 89.9 -131.7 cm. Mean ears per plant was 1.14 with ranges of 0.79 -1.48. Anthesis-silking interval ranged from -0.2 to 3.3 d with a mean of 1.6 d. Hybrid x E, GCA x E and SCA x E interactions were highly significant for all traits except SCA x E for ear height and anthesis-silking interval.
HAOM=Harare optimal, RAOM=Rattray optimal, MPOM=Mpongwe optimal, BKOM=Bako optimal, MLOM=Melkasa optimal, PWOM=Pawe optimal, AWOM=Awassa optimal, JMOM=Jimma optimal, CHDS=Chiredzi stress, KBDS=Kiboko stress, KBOM=Kiboko optimal. ‡ proportion of QPM hybrid with higher grain yield than the best check (normal maize or QPM); SE(M)= standard error of the mean   Table 5. Mean squares from combined analysis of variance and means for grain yield and agronomic traits of QPM hybrids across nine optimal environments, 2006 -2008 Estimates of GCA effects for grain yield showed that inbred lines VL05561, VL05483, CML511, CML159 and VL06375 combined well in most of the environments (Table 6). These inbred lines mostly showed positive and highly significant GCA effects in most environments. On the other hand, VL052, VL052887, VL0523 and CML144 showed negative and highly significant GCA effects in most of the environments. Inbred lines VL05561, VL05483 and CML511 showed high positive GCA effects across optimum and combined environments. For days to anthesis, VL054178, VL05482, VL05561, VL05483, CML511 and VL06375 had negative and highly significant GCA effects in most environments (Table 7). On the other hand, inbred lines VL05200, VL054178, VL052887, VL0523, VL05561 and CML144 showed positive and highly significant GCA effects in most environments. VL054178, VL05482, VL05561, VL05483, CML511, CML159 and VL06375 had highly significant negative GCA effects for days to silking for both water stress and optimal environments. Inbred lines VL054178, VL05482, VL05561, VL05483 and VL06375 had negative and highly significant GCA effects for days to silking (Table 8). On the other hand, VL05468, VL052887, VL0523, VL0524, CML144 and CML491 showed positive and highly significant GCA effects. VL054178, VL05482, VL05561, VL05483, CML159 and VL06375 had highly significant negative GCA effects for days to anthesis for both water stress and optimal environments. The GCA effects for anthesis-silking interval were negative and highly significant for VL05561 but positive and highly significant for VL054178 in almost all environments (Table  9). Across water stress environments, inbred lines VL054178 and VL05482 showed lower GCA effects. VL052887, VL05561 and CML144 had negative and highly significant GCA effects across optimal environments. VL054178, VL05561, VL05483 and VL06375 showed lower GCA effects for anthesis-silking interval over all environments. Inbred lines VL05200, VL054178, VL05482, CML144 and CML159 showed negative and significant GCA effects for plant and ear height in most environments (Tables 10 and 11). However, VL05483 and VL06375 had positive and significant GCA effects for plant height while VL053, VL0524 and VL5561 had positive and significant GCA effects for ear height in most environments. For ears per plant, inbred lines VL05482, VL05483 and CML511 showed positive and significant and VL05200, VL05468, VL0523, VL0524 and CML159 showed negative and highly significant GCA effects in water stress and optimal environments (Table 12). At Chirezi under water stress, VL05482, CML511 and CML491 showed negative and significant GCA effects.

Discussion
The results observed in various environments (Table 2) showed that water stress significantly affected grain yield, as previously reported Edmeades, 1993, 1996;Banziger et al., 1999a;Derera et al., 2008). High levels of variation observed among hybrids under water stress, and optimal environments indicate the possibility of selecting for improved grain yield and agronomic traits under stress and non-stress conditions. The existence of genetic variability in maize evaluated under stress conditions has been reported by several investigators Bolanos and Edmeades, 1996;Beck et al., 1997;1999b;Betran et al., 2003;Derera et al., 2008). Significant GCA and SCA mean squares for most traits in each environment indicate the importance of both additive and non-additive effects for the traits studied. This suggests that effective selection or systematic hybridization could be employed in improving these traits. HAOM=Harare optimal, RAOM=Rattray optimal, MPOM=Mpongwe optimal, BKOM=Bako optimal, MLOM=Melkasa optimal, PWOM=Pawe optimal, AWOM=Awassa optimal, JMOM=Jimma optimal, CHDS=Chiredzi stress, KBDS=Kiboko stress, KBOM=Kiboko optimal; * P≤ 0.05 ; ** P≤ 0. HAOM=Harare optimal, RAOM=Rattray optimal, MPOM=Mpongwe optimal, BKOM=Bako optimal, MLOM=Melkasa optimal, PWOM=Pawe optimal, AWOM=Awassa optimal, JMOM=Jimma optimal, CHDS=Chiredzi stress, KBDS=Kiboko stress, KBOM=Kiboko optimal; * P≤ 0.05 ; ** P≤ 0.01; ‡ ACDRT= across water stress environments; # ACOPT= across optimum environments; P1= VL052; P2= VL05200; P3= VL05468; P4= VL054178; P5= VL052887; P6= VL05482; P7= VL0523; P8= VL0524; P9= VL05561; P10= VL05483; P11= CML511; P12= CML144; P13= CML159; P14= CML491; P15= VL06375; SE(gi)= standard error of GCA effects Combined analysis of variance across water stress (Table 4) and optimal (Table 5) environments indicated the existence of significant variation among hybrids and environments for all traits. Both additive and non-additive genetic effects were not important for grain yield across water stress environments while only additive effect was important for days to anthesis and silking, anthesis-silking interval and ears per plant. This finding is contrary to the reports of other researchers (Betran et al., 1999;Makumbi et al., 2004;Derera et al., 2008), who reported the importance of additive effects for grain yield of normal maize under water stress. When genetic variance for grain yield is not apparent, secondary traits of adaptive value whose genetic variability increases and whose heritability remains high under water stress can increase selection efficiency (Bolanos and Edmeades, 1996;Edmeades et al., 1997;Banziger et al., 1999b). Highly significant GCA and SCA mean squares for all traits under optimal environments indicate the importance of both additive and non-additive gene effects for the inheritance of these traits. Similar results have been reported in diallel studies of QPM inbred lines under optimal environments (Pixley and Bjarnason, 1993;Bhatnagar et al., 2004;Hadji, 2004;Fan et al., 2004). Derera et al. (2008) reported the importance of both additive and non-additive effects in conditioning grain yield, days to anthesis and silking, and anthesis-silking interval in Design-II crosses of normal maize inbred lines. Similarly, additive and non-additive effects were important for all traits evaluated across environments except anthesis silking interval which had non-significant SCA effects. Significant mean squares of Hybrid x E, GCA x E and SCA x E interactions for most traits across environments indicate that these effects were not consistent over environments. This implies that different genes are involved in controlling these traits under water stress and optimal conditions. Cooper and Byth (1996) explained that the larger the degree of genotype-by-environment interaction, the more dissimilar the genetic systems controlling the physiological processes conferring adaptation to different environments. Even though significant cross-over interactions were observed for GCA effects of the inbred lines, some inbred lines were identified with consistent GCA effects across environments. This implies that the genetic systems controlling a given trait under different stress and nonstress conditions are at least partially similar. Hence, it is possible to identify QPM hybrids that perform well across stress levels in Africa. Similar conclusions have been drawn by Betran et al. (2003) who evaluated tropical normal maize inbred lines and their hybrids for grain yield under optimal and water stress conditions. Inbred lines VL054178, VL05561, VL05483, CML511, CML159 and VL06375 were good general combiners for grain yield in both water stress and optimal environments indicating that these inbred lines contributed to increased grain yield in their crosses under all environmental conditions. Inbred lines VL054178, VL05482, VL05561, VL05483, CML159 and VL06375 contributed to earliness under most environments as inferred from the negative and highly significant GCA effects of days to anthesis and silking. VL05561 was the best general combiner for anthesis-silking interval. Inbred lines VL05200, VL054178, VL05482, CML144 and CML159 were good combiners for plant stature as they contributed to reduced plant and ear height in the crosses. VL05482, VL05483 and CML511 contributed to increased ears per plant in the crosses. Anthesis-silking interval and ears per plant are important secondary traits to be considered in increasing the efficiency of selection for grain yield under stress. The highest grain yielding genotypes under water stress tended to show lower anthesis-silking interval, delayed senescence, and a higher number of ears per plant Banziger et al., 1999c;Diallo et al., 2004). Higher SCA variances than GCA variances for grain yield in most optimal environments indicate that additive variability was of greater importance in the inheritance of grain yield under optimal conditions. Under water stress conditions, however, additive variability was more important than non-additive variability. The predominance of additive effects under water stress conditions has been reported by several researchers (Betran et al., 2003;Diallo et al., 2003;Makumbi et al., 2004;Derera et al., 2008). Additive effects were more important that non-additive effects in the inheritance of days to anthesis and silking in all cases. Similarly, additive effects were more important for anthesissilking interval, plant and ear height, and ears per plant in most cases. According to Baker (1978), when SCA mean squares are not significant, the hypothesis that the performance of a single-cross progeny can be adequately predicted on the basis of GCA would be accepted. On the other hand, if the SCA mean squares are significant, the relative importance of GCA and SCA should be assessed by estimating components of variance in determining progeny performance.

Conclusions
A large proportion of the maize crop in Africa is grown by small scale farmers under low input systems, without adequate fertilization and irrigation. Significant yield losses due to water stress were realized in this study. The results indicated the availability of considerable variation among QPM hybrids and the possibility of making selections for grain yield and agronomic traits under stress and non-stress conditions. Significant GCA and SCA mean squares, and hence the importance of both additive and non-additive effects was observed for most traits in most environments. Neither additive nor non-additive genetic effects were important for grain yield across water stress environments. In this case, secondary traits such as anthesis-silking interval and ears per plant with high genetic variability and heritability can be used to increase selection efficiency. Estimates of GCA effects showed that inbred lines VL054178, VL05482, VL05561, VL05483, CML511, CML159, CML491 and VL06375 had good GCA effects for most traits under stress and non-stress conditions. These inbred lines can be used for the development of QPM hybrids and synthetics that perform well across stress and non-stress environments. In general, the inbred lines used in this study were found to be useful sources for genetic variability for the development of new genotypes for stress tolerance and the study confirmed the possibility of achieving good performances across stress and non-stress conditions in QPM germplasm.