|
|
||||||||
a Dep. of Agronomy, Iowa State Univ., Ames, IA 50011
b CIMMYT, Apdo Postal 6-641, 06600, Mexico D.F., Mexico
c Deceased 2004,, Midlands State Univ., Faculty of Natural Resources Management and Agric., P.B. 9055, Gweru, Zimbabwe
* Corresponding author (k.pixley{at}cgiar.org)
| ABSTRACT |
|---|
|
|
|---|
Abbreviations: CIMMYT, International Maize and Wheat Improvement Center CML, CIMMYT maize line CTH, controlled temperature and humidity GCA, general combining ability masl, meters above sea level MDP, median development period SCA, specific combining ability
| INTRODUCTION |
|---|
|
|
|---|
Weevil damage results directly in lost food (reduced grain weight), and may also reduce future maize production for farmers who plant saved grain as seed, a practice that accounts for about 70% of all maize planted in eastern and southern Africa (Pingali and Pandey, 2001). There also may be a health risk associated with consuming weevil-infested maize grain, as it has been reported to commonly have higher levels of Aspergillus flavus Link:Fr. (a fungus that produces aflatoxins) contamination than noninfested maize kernels (Smalley, 1998). For all of these reasons, weevil-resistant maize varieties would be a valuable component of integrated storage pest management strategies for maize.
Painter (1968) described three mechanisms of resistance to insect pests: nonpreference (for oviposition, food, or shelter), antibiosis (adverse effect of plant on biology of pest), and tolerance (repair, recovery, or ability to withstand infestation). Although genetic variation for weevil nonpreference (evaluation of resistance in a free-choice environment) has been reported among maize cultivars (Kang et al., 1995; Derera et al., 2001b), these findings are of doubtful utility to farmers because most store only one or very few varieties together in any storage structure, thereby giving weevils little or no choice. Tolerance to weevil damage might be of some interest if it involves resistance to secondary pathogens or compensation mechanisms that ensure early seedling vigor, but tolerance to weevil damage is unlikely to compensate for grain weight loss through weevil tunneling and feeding. It is therefore antibiosis, or resistance expressed in a no-choice test, that is of greatest relevance to farmers.
Variability for antibiosis resistance to maize weevil exists in maize (Giga and Mazarura, 1991; Derera et al., 1999), and there are a few published reports about the inheritance of resistance to maize weevil (Derera et al., 2001a; Kang et al., 1995; Tipping et al., 1989; Widstrom et al., 1975). Widstrom et al. (1975) used a North Carolina Design II with 80 inbred lines and reported a significant maternal effect for weevil resistance in F1 single-cross and topcross hybrids in no-choice experiments. They also reported a large intralocus nonadditive genetic effect at the maternal level, suggesting importance of dominance effects of the maternal genotype for resistance to maize weevil. Derera et al. (2001a) evaluated a North Carolina Design II mating among 18 maize inbred lines and reported greater importance of female parent GCA than male parent GCA in both F1 and F2 grain, indicating that maternal effects were important for weevil resistance in no-choice experiments. Tipping et al. (1989) studied the number of weevil eggs laid in no-choice experiments for grain of each genotype for a 10-parent diallel mating design and reported GCA and to a lesser extent SCA as important; significant maternal effect was detected in the F1, but was not significant in F2 grain. Kang et al. (1995) found significant GCA, SCA, and maternal effects for grain weight loss caused by weevils for a 10-parent diallel set of hybrids evaluated in a free-choice environment. Also in free-choice experiments, Derera et al. (2001b) found female and male parent GCA were of similar importance for weevil resistance. Except for Derera et al. (2001a)(2001b), most of the germplasm used in inheritance studies for resistance to maize weevil cited herein were conducted with temperate maize inbred lines.
Our objectives were to (i) estimate combining ability effects for resistance to maize weevil and grain yield among 14 inbred lines adapted to midaltitude environments of southern Africa, (ii) assess the importance of reciprocal effects for weevil resistance of maize hybrids in F2 and F2Syn 1 grain, and (iii) determine the relationship between grain yield and weevil resistance to assess the feasibility of improving both.
| MATERIALS AND METHODS |
|---|
|
|
|---|
|
53000 plants ha1. Plots were fertilized with 32 kg N, 56 kg P2O5, 28 kg K2O, and 3.2 kg Zn ha1 at planting and supplemented with 69 kg N ha1 6 wk after planting. F2 grain was formed by self-pollinating 10 to 20 F1 plants within a row. Ears from each row were placed in mesh bags and dried to about 10% moisture content by blowing warm air (<36°C) in a forced-air drier at Harare. Ears were shelled in bulk and the grain was kept in a cold room maintained between 5 and 8°C, without chemical treatment. F2 grain was planted and recombined to obtain F2Syn 1 (terminology of Lamkey et al., 1995) at Harare during the summer 19992000 by intermating five to eight plants (full-sib mating, controlled hand pollination) within single, 4-m rows for each F2. Nursery management was as described for Muzarabani, except that the nursery at Harare was mainly rain fed. Irrigation was used as necessary to ensure that neither generation was exposed to severe moisture stress during growth. F2Syn 1 ears were dried, shelled in bulk, and grain was stored following the same procedures as described for F2 grain.
Weevil Resistance Evaluation
Experiment 1: F2 Grain of a 14-Parent Diallel
F2 grain from all hybrids was evaluated for resistance to maize weevil during 2000 in a controlled temperature (28 ± 2°C) and humidity (70 ± 5% RH) room (CTH room) at CIMMYT, Harare. Two hundred grams of clean, undamaged grain from each of the 182 single crosses was weighed into a 1000-mL jar with a brass screen lid that allowed adequate ventilation, and placed in a freezer at 20°C to destroy any insects or eggs that could have been present in the grain. The grain samples were then placed in the CTH room for a 3-wk conditioning period to achieve uniform grain temperature and moisture content. Four subsamples (replicates) of 50 g were weighed into 500-mL jars, infested with 32 unsexed weevils aged 7 to 14 d, and placed in the CTH room for a 10-d oviposition period. The weevils were then removed and the numbers of dead and living weevils were used to calculate percentage parent mortality. The samples were immediately returned to the CTH room for 21 d, after which the number of F1 weevils (progeny) that emerged from each sample was counted every 2 d until no more weevils emerged. The median development period (MDP) was calculated for each maize genotype as the time in days by which 50% of the weevils emerged (hatched). Dobie index of susceptibility (Dobie, 1977) was calculated as 100 x [loge (total number of progeny emerged)/MDP]. Both Dobie index of susceptibility and total progeny emerged were used as resistance parameters and subjected to ANOVA.
Experiment 2: F2 and F2Syn 1 Grain of a 10-Parent Diallel (Subset of Experiment 1)
Assessment of weevil resistance of F2 grain was repeated in 2002, together with evaluation of the F2Syn 1 grain. However, because of small grain quantities for some F2 and F2Syn 1 crosses, only 10 parents (Table 1) were included in the final diallel analyses, and samples were replicated three times for this evaluation. The same procedure described above for 2000 was used to evaluate weevil resistance in 2002, except that weevils were counted only once (at the end of a 45-d incubation period). Therefore, Dobie index of susceptibility was not calculated in 2002, and we relied on the number of weevils emerged from each sample as the resistance parameter for this data set. Use of total number of weevils emerged at the end of the incubation period instead of Dobie index of susceptibility is justified by the huge savings in labor, increased number of grain samples that can be evaluated in the research program, and by numerous data sets that indicate the correlation between Dobie index and number of weevils emerged is typically >0.90 (e.g., Classen et al., 1990; Arnason et al., 1994).
Yield Trials
Experiment 3: F1 Hybrids of the 14-Parent Diallel (Experiment 1)
The yield trial of the F1 hybrids from the 14-parent diallel was conducted at four sites in Zimbabwe: Harare (17°49' S lat., 1506 masl), Glendale (17°5' S lat., 1200 masl), Agricultural Research Trust Farm (17°26' S lat., 1480 masl), and Rattray Arnold Research Station (17°40' S lat., 1308 masl) during the 19992000 summer (NovemberApril). All four locations have deep reddish-brown granular kaolinitic clay Rhodustalf soils. Reciprocal F1 crosses were bulked and planted using an
-lattice design (Patterson et al., 1978) with two replications at each site. Plots were two rows, 4 m long, with 0.75 m between rows at all sites except Harare, where single-row plots were used, but with a final plant population of
53000 plants ha1 at all four sites. Recommended fertilizer rates were used for each location. Yield was shelled grain adjusted to 12.5% moisture content.
Statistical Analyses
Griffing's (1956) Method 3 (hybrids and reciprocals, without parents) Model 1 (fixed effects model) was used to analyze weevil resistance data (log-transformed number of progeny emerged and Dobie index of susceptibility) for the 14-parent diallel (Exp. 1). Combining ability analyses for the F2 grain of the 14-parent (Exp. 1), and for both F2 and F2Syn 1 grain of the 10-parent diallel (Exp. 2) used the DIALLEL-SAS program (Zhang and Kang, 1997), which allowed partitioning of reciprocal effects into maternal (general reciprocal) and nonmaternal (specific reciprocal) components as suggested by Cockerham (1963) and demonstrated by Pederson and Windham (1992) and Kang et al. (1999). In these analyses, GCA and SCA are calculated as described by Griffing (1956), using the two reciprocal crosses for each hybrid. Reciprocal effects are orthogonally subdivided into those due to maternal effects, which are the average effects of using parents as females rather than males in their hybrids, and nonmaternal, which are specific reciprocal effects or deviations from the maternal effects. Genetic expectations for diallel analysis of F2 and F2Syn 1 grain are same as for F1 hybrids; however, due to decreased heterozygosity, the dominance contribution to SCA effects is theoretically half for F2 and F2Syn 1 as compared with evaluation of F1 generation (Hill et al., 2001).
Data for the 10-parent subset diallel were combined across weevil resistance evaluation dates (Exp. 1 and 2) in an ANOVA using DIALLEL-SAS via the GLM (general linear model) procedure of SAS (SAS Institute, 2001). Hybrid and combining ability effects were considered fixed, while the evaluation date effect was considered random. Main effects such as hybrids and their subclasses were tested for significance using their respective interaction terms with experiments (evaluation dates) as the error terms, while the interaction variances were tested for significance using the pooled error.
Grain yield was analyzed using Griffing's (1956) Method 4 Model 1 for individual sites (data not shown) and across sites, for both the 14-parent and subset 10-parent diallels. Environments were considered random and genotypes as fixed effects.
| RESULTS AND DISCUSSION |
|---|
|
|
|---|
Hybrid, GCA, SCA, and reciprocal effects were highly significant (P < 0.01) for weevil progeny emerged for F2 grain of the 14-parent diallel (Table 2). Partitioning of the reciprocal sums of squares indicated that both maternal (general reciprocal) and nonmaternal (specific reciprocal) effects were highly significant (P < 0.01). Maternal effect (general reciprocal) is the effect of maternal genotype or tissue on a trait in its offspring, whereas nonmaternal effect (specific reciprocal) is the effect of extranuclear genetic factors or interaction between nuclear and plasma genes (Kang et al., 1999; Zhang and Kang, 1997).
|
|
Combined analysis across evaluation experiments revealed that P1 and P2 (see Table 1) had significant positive GCA effect for number of weevil progeny emerged, and hence contributed weevil susceptibility to their hybrids (Table 4). The positive GCA effect was also significant for both of these lines in the full 14-parent diallel and was significant even for F2Syn 1 grain of hybrids of P2. By contrast, GCA effects were significant and negative for CML387, indicating that it contributed to weevil resistance of its hybrids. CML387 seems particularly useful, considering that it also had good combining ability for grain yield (Table 4).
|
Because of the important role of pericarp, which is maternal tissue, in protecting the maize kernel from weevil attack (Serratos et al., 1993), we expected lines with best GCA to also have favorable maternal effect for weevil resistance parameters. This was not always the case, however, as illustrated by both P2 and P8. Although we cannot explain this result, we note that a similar result was reported by Kang et al. (1995), who measured negative GCA and positive maternal effect for grain weight remaining after infestation by maize weevil for 3 mo for the temperate inbred maize line Mo17. They recommended that Mo17 should be used as a female in hybrids for weevil-prone environments, despite being susceptible (negative GCA) to weevil.
The GCA, SCA, and reciprocal effects accounted for 26, 45, and 29%, respectively, of hybrid sums of squares (Exp. 2), and F tests confirmed that SCA were more important than GCA and reciprocal effects for weevil progeny emerged from F2 grain (Table 3). The GCA and SCA effects accounted for similar portions of hybrid sums of squares (39 and 36%, respectively), while reciprocal effects were not significant for weevil progeny emerged from F2Syn 1 grain. Thus, even though additive gene action was important for determining weevil resistance of F2 and F2Syn 1 grain, significant variance caused by nonadditive effects indicates that weevil resistance of hybrids could not be accurately predicted based only on their parents' resistance (Table 1). Clearly, inheritance of weevil resistance is complex and heritability is likely small to moderate. Dhliwayo and Pixley (2003) have demonstrated, however, that weevil resistance is sufficiently heritable to enable successful improvement in maize.
Grain Yield
Variance for grain yield among hybrids was highly significant (P < 0.01) for both the 14-parent and the 10-parent subset diallels (data not shown). For the 14-parent diallel, both GCA and SCA effects were significant (P < 0.01), whereas only GCA effects were significant (P < 0.01) for grain yield for the 10-parent diallel. We did not necessarily expect large SCA effects for grain yield in this study because selection of the inbred parents was based on their resistance to maize weevil, and not their heterotic behavior for grain yield. Mean grain yield across all four locations for the 10-parent diallel ranged from 3.36 for P2 x P6, to 6.77 Mg ha1 for P4 x P6.
Associations among Traits and Combining Ability Effects
Weevil resistance of F2 and F2Syn 1 grain was moderately consistent as indicated by phenotypic correlation coefficient of r = 0.52 (P < 0.01) between weevil progeny number emerged from grain of the two generations. The GCA effects of parents for weevil resistance of F2 and F2Syn 1 grain were significantly correlated (r = 0.72, P < 0.05), confirming that additive effects were important, and providing further evidence of consistency between weevil resistance of F2 and F2Syn 1 grain. The SCA effects for progeny emerged from grain of F2 and F2Syn 1 hybrids were not significantly correlated.
Grain yield of F1 hybrids was not significantly associated with weevil progeny number emerged from F2 (r = 0.23) or F2Syn 1 grain (r = 0.22). The GCA effects for grain yield were not correlated with GCA effects for weevil progeny emerged from F2 (r = 0.20) or F2Syn 1 grain (r = 0.21). Thus, weevil resistance and grain yield were not significantly associated with each other in this study.
| CONCLUSIONS |
|---|
|
|
|---|
Because some of the inbreds used in this study are elite germplasm, and are already widely used in maize breeding programs, there is scope for producing hybrids with improved levels of resistance to maize weevil. The lack of correlation between grain yield and resistance to maize weevil is important because it indicates that both of these traits can be improved simultaneously. Lines such as CML387, with positive GCA for grain yield and negative GCA for number of weevils emerged following infestation, should be useful for producing superior hybrids or populations.
| ACKNOWLEDGMENTS |
|---|
Received for publication December 15, 2003.
| REFERENCES |
|---|
|
|
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Vadose Zone Journal | |||
| Journal of Plant Registrations | Soil Science Society of America Journal | ||||
| Journal of Natural Resources and Life Sciences Education |
Journal of Environmental Quality |
||||