Assessing performance of white endosperm testers with varying resistance reactions to Striga ( Striga hermonthica ) for evaluating resistant maize ( Zea mays ) inbred lines

Identification of testers is crucial for hybrid maize breeding programme. However, limited information is available about ideal testers for characterising the combining ability of Striga resistant maize inbreds. This study was conducted to assess the relative value of three inbred testers with varying resistance reactions to Striga for determining the combining ability of Striga resistant inbreds. Ninety testcrosses involving 30 Striga resistant inbreds and three testers were evaluated under artificial and natural Striga infestation and non-infested conditions at two locations for 2 years. Lines x tester interaction was significant ( p ≤ .05) for most traits, indicating differential ranking of lines by the testers. The GCA effects of testers for most traits were high, highlighting the predominance of additive gene action in controlling the overall performance of testcrosses. The resistant and tolerant testers exhibited desirable GCA effects, broader testcross performance, greater genetic variances and consistent ranking of testcrosses under both growing conditions than the susceptible tester. These testers can be successfully used for identifying superior Striga resistant inbreds to develop high yielding and resistant hybrids for commercialization.


| INTRODUCTION
In a relatively short period of time after introduction, maize has become a major staple cereal crop across regions in sub-Saharan Africa currently occupying 40 percent of the total cereal production area in the continent (Ekpa et al., 2019).The demand for maize in African and other developing countries will double by 2050 (CIMMYT IITA, 2010).Increasing the productivity of maize in smallholder farmers' fields will contribute to meeting the increasing demand for food and other uses.This may be achieved by supporting resilience of smallholder farming communities through access to technologies to mitigate the adverse effects of climate change and prevalent pests and diseases including a parasitic weed known as Striga hermonthica that perpetuate low grain yields of most cereals (Cairns et al., 2012;Gichuru, 2013).S. hermonthica depletes nutrients and water from the host plants such as maize thereby affecting the productivity of the crop (Kim, 1991;Spallek et al., 2013) in millions hectares of infested land across Africa (Lagoke et al., 1991).
The development and deployment of maize varieties with polygenic resistance to Striga has thus been considered a critical component of an integrated control strategy to minimise yield losses in farmers' fields as most of the maize cultivars currently grown by farmers may suffer close to 100% yield loss under severe Striga infestation (Kling et al., 2000;Menkir et al., 2007).Maize breeders at the International Institute of Tropical Agricultural (IITA) made significant advances in identifying tolerant white maize germplasm that were used for developing diverse source populations (Kim, 1991;Kim & Winslow, 1991).The resulting diverse populations have been sources of resistant maize inbred lines with desirable agronomic traits, supporting fewer emerged parasites and sustaining less Striga damage symptoms to develop resistant hybrids targeted to areas infested with the parasite (Menkir et al., 2016;Menkir & Meseka, 2019).
The success of resistance hybrid maize breeding programme depends on the choice of the most appropriate tester that correctly determine the combining abilities of Striga resistant maize inbred lines to select potential parents of superior hybrids for commercialization (Hallauer & Lopez-Perez, 1979).As testers with high frequency of favourable alleles have been recommended to identify lines that form hybrids with superior specific combining ability (Hallauer & Carena, 2009), breeders in IITA have used Striga resistant white inbred testers to assess the performance of new inbred lines in hybrid combinations under artificial Striga infestation.However, the potential usefulness of other testers with susceptible and tolerant reactions to the parasite in determining the combining ability of resistant inbred lines has rarely been reported.
Several studies have been conducted on the identification of suitable testers in maize breeding programme s providing different recommendation.Many recommend use of testers that are simple to use and maximise the genetic difference expressed among testcrosses of inbred lines in hybrid breeding (Allison & Curnow, 1966;Hallauer et al., 2010;Hallauer & Miranda, 1988;Matzinger, 1953;Miranda et al., 2012;Rawlings & Thompson, 1962;Russell, 1961).Moreover, lines and populations with low frequency of favourable alleles have been regarded as desirable testers in most breeding programmes because they permit easier identification of lines with high frequency of favourable allele for target traits (Hallauer et al., 2010;Hallauer & Miranda Filho, 1995).In contrast, Hallauer and Carena (2009) suggested the use of elite inbred lines with high frequencies of favourable alleles as testers to identify superior hybrids for direct commercialization.The differing recommendations highlight the need for further studies to identify and use testers particularly in Striga resistance hybrid breeding programmes.
The line x tester method of analysis has been extensively used by breeders to evaluate the combining abilities of lines to select suitable parents for hybrid formation (Izhar & Chakraborty, 2013;Kamara et al., 2014;Legesse et al., 2009;Menkir et al., 2004;Rahman et al., 2013;Ruswandi et al., 2015).This method has proven useful to identify appropriate testers for use in breeding programmes (Gutierrez-Gaitan et al., 1986;Li et al., 2007;Vasal, Srinivasan, Han, & Gonzales, 1992a).The fact that identification of desirable testers and their discriminating capacity is still an unsolved issue in crop breeding (Hallauer et al., 2010), evaluating the relative importance of using different types of testers will be useful to select those that are suitable for resistance breeding programs (Gutierrez-Gaitan et al., 1986;Li et al., 2007;Vasal, Srinivasan, Pandey, et al., 1992b).Also, conducting continual studies to identify new testers will be important for improving the efficiency of discriminating among new maize inbred lines emanating from a resistance breeding programme (Guimaraes et al., 2012).Zebire et al. (2020) found that Striga resistant followed by Striga tolerant yellow maize inbred lines were suitable testers under artificial Striga infested and non-infested conditions.However, the relative superiority of these testers has not been confirmed under natural field infestation.The present study was, therefore, conducted to determine the usefulness of white inbred lines with varying resistance reactions to S. hermonthica to select new Striga-resistant maize inbred lines with superior hybrid combinations under artificial, natural Striga infestation and non-infested conditions.

| Genetic materials
A total of 30 Striga resistant white maize (Zea mays) inbred lines and three testers with varying levels of resistance to S. hermonthica (Del.) Benth developed at IITA were used for this study (Table S1).The 30 inbred lines were extracted from bi-parental crosses of pairs of Striga resistant lines as well as synthetics, broad-based populations and a backcross containing Zea diploperennis (Z.diplo) as a donor of favourable resistance alleles (Kling et al., 2000).Repeated screening of selected lines from each of the synthetics, broad-based populations and a backcross containing Zea diploperennis under artificial Striga infestation, both in the screen house and in the field led to the development of the 30 Striga-resistant maize inbred lines (Menkir, 2006).
The three testers representing three heterotic groups were also selected from diverse source populations with different genetic backgrounds (Table S1).The first tester (T1) is a Striga resistant tester derived from the backcross population containing Zea diploperennis as a donor of resistance to the parasite, while the second tester (T2) is a Striga tolerant tester derived from a cross of a tropical population with a temperate inbred line (N28/TZSR).The third tester is a Striga susceptible tester (T3) derived from a cross of two tropical populations (TLATT7844/TLSR).The genetic backgrounds of the sources of the resistant maize inbred lines and testers have been fully described by Kling et al. (2000) and Kim et al. (1998).

| Generation and evaluation of testcrosses
The 30 maize inbred lines each were crossed to the three testers to generate 90 testcrosses under irrigation during the 2015/2016 dry season at IITA in Ibadan, Nigeria.The 90 testcrosses along with known tolerant and susceptible benchmark hybrids were evaluated under artificial Striga infestation and non-infested conditions in Abuja and Mokwa in 2017 and 2018 as well as under natural field infestation at Agwanmalamai and Samaru in 2017 and 2018.In each location, the 92 hybrids were arranged in a 4 Â 23 alpha lattice design with two replications.In Abuja and Mokwa, the hybrids were planted in a criss-cross arrangement (Pearce, 1976) that allowed planting of each hybrid in adjacent infested and non-infested strips, which were located directly opposite to each other and separated by a 1.5-m alley.
This arrangement provides precise estimates of yield loss attributable to S. hermonthica damage (Kling et al., 2000).In the current study, the procedure described by Menkir and Meseka (2019) was followed to inoculate the infested strip and keep the non-infested strip free from Striga seeds.Within each strip, a hybrid was planted in a 4-m-long row with an inter-row spacing of .75m and an intra-row spacing of .25 m.In addition, the trials at Agwanmalamai and Samaru were planted under natural field infestation using the hot spots for Striga infestation.In all the trials, two maize seeds were placed in the same hole with the Striga seeds and later thinned to one maize plant per hill at 2 weeks after planting (WAP).Fertilizer was applied at the rate of 30 kg/ha each of N, P and K at 30 days after planting.An additional 20 kg N ha À1 was applied as urea at 6 WAP.This delay in application and reduced rate of fertilizer was necessary to enhance the germination of Striga seeds and subsequent attachment of Striga plants to the roots of the host plants in Striga-infested plots (Kim, 1991).Weeds other than Striga were controlled by hand weeding in both infested and non-infested fields.

| Trait measurements
Data recorded in each plot under all research conditions included plant stand which was counted as the total number of plants per plot obtained immediately after thinning.Days to anthesis and silking were recorded as the number of days from planting to when 50% of the plants in a plot had anthers shedding pollen and showing emerged silks, respectively.Anthesis-silking interval was calculated as interval in days between dates of silking and anthesis.Plant height was measured in cm as the distance from the base of the plant to the height of the first tassel branch.Ear aspect was scored on a 1 to 9 scale, where 1 = clean, uniform and large ears, and 9 = rotten, variable and small ears.All ears harvested from each plot were shelled to determine percent moisture, which was used to determine grain yield adjusted to 15% moisture under both infested and non-infested conditions.Host plant damage symptoms were visually rated in each infested row at 8 and 10 weeks after planting using a scale of 1 to 9, where 1 = no visible host plant damage symptom and 9 = all leaves completely scorched, resulting in premature death (Kim, 1994).Also, the total numbers of emerged S. hermonthica plants were counted in each infested row at 8 and 10 weeks and were divided by the corresponding plant stand to obtain the number of emerged parasites per plant in each row.The total number of plants and ears were counted in each Striga-infested plot at the time of harvest and used to calculate the number of ears per plant.Ear height was measured in cm as the distance from the base of the plant to the height of the node bearing the upper ear.Husk cover was rated on a 1 to 5 scale under non-infested condition, where 1 = husks tightly arranged and extended beyond the ear tip and 5 = ear tips exposed.Plant aspect was rated on a scale of 1 to 9 in non-infested plots, where 1 = excellent plant type with large and similar ears, low ear placement, shorter plants, resistance to foliar diseases and little stalk and root lodging, and 9 = plants with small and variable ears, high ear placement, tall plants, susceptible to foliar diseases as well as stalk and root lodging.Also, ear rot was rated on a scale of 1 to 5, where 1 = little or no visible rotting of the ears and 5 = extensive visible rotting of the ears.

| Statistical analysis
Each location-year combination in the current study was regarded as test environment to conduct analyses of variance for all traits measured under infested and non-infested conditions using PROC MIXED procedure in SAS (SAS, 2013).In these analyses, environments, replication (environments), block (replication Â environments), were considered as random effects, whereas testcrosses were regarded as fixed effects.Further line x tester analysis was conducted to partition the testcross mean square into lines, testers, line x tester, environment x line, environment x tester and environment x line x tester effects using an SAS program following the procedure of Singh and Chaudhary (1977).The efficiency of testers was determined based on the genetic variance estimates derived from combined analysis of variance of testcrosses means of each tester obtained from the four environments (Castellanos et al., 1998).Least significance test (LSD = .05)was computed to compare mean differences between testers for each trait.Second, the testcross means obtained in each of the four test environments were ranked using PROC RANK in SAS (SAS, 2013).The resulting ranks were then used to calculate Kendall's coefficient of concordance (W) to assess the similarity of the rank order of the 30 testcrosses across the four test environments (Kendall & Smith, 1939).The significance of GCA and SCA effects were tested by dividing the corresponding GCA and SCA values by their respective standard errors and comparing the resulting observed t-value with tabular t-value using SAS.Baker's (1978) ratio of GCA: SCA mean square was computed to determine whether additive or non-additive gene effect determine the inheritance of each trait.

| Variations among testcrosses under Strigainfested and non-infested conditions
Environment and testcrosses effects were significant (p < .05)for all the traits recorded under Striga infestation.Testcross Â environment interaction was significant (p < .01)for all traits except for days to anthesis, plant height, and number of ears per plant and Striga damage rating at 8 and 10 WAP (Table 1).The GCA effects due to lines and testers were significant (p < .01)for most of the traits.Line Â tester interactions (SCA) were also significant (p < .01)for grain yield, anthesis-silking interval, ears aspect, plant height and number of ears per plant.Line Â environment interactions were significant (p < .01)for all traits except for days to anthesis, while tester Â environment interaction was significant (p < .01)for grain yield, days to silking, anthesis-silking interval, ear aspect, plant height and number of ears  per plant.Line Â tester Â environment interaction was not significant for any of the traits measured under infestation (Table 1).
Under natural field infestation, environment and testcrosses effects were significant (p < .05)for all the traits measured under Striga infestation except for Striga damage rating at 8WAP (Table 2).
Testcross Â environment interaction was only significant (p < .01)for plant height, ear aspect and number of ears per plant.The GCA effects due to lines and testers were also significant (p < .01)for most of the traits, whereas line Â tester interactions (SCA) were significant (p < .01)for grain yield and five other traits.Both line x environment and tester x environment interactions were not significant for grain yield and most other traits.Again, line Â tester Â environment interaction was not significant for all the traits (Table 2).Under noninfested condition, environment and testcross mean squares were significant (p ≤ .01)for all the traits except for number of ears per plant for testcrosses (Table 3).Testcross Â environment interaction was only significant for ear aspect, number of ears per plant and plant aspect.Line and tester GCA effects were significant for all the traits, except for number of ears per plant for lines and days to anthesis for testers.Line Â tester interaction was significant for grain yield and five other traits.Line Â environment interaction was significant (p < .001)for all traits except days to anthesis, plant height, ear height and husk cover, while tester Â environment interaction was only significant for grain yield, ear aspect, and plant height and plant aspect.
The line Â tester Â environment interaction was significant only for ear aspect and number of ears per plant (Table 3).
The genetic variances estimated for the testcrosses of the resistant tester (T1) were the highest for most of the traits measured under Striga infestations, while the genetic variances of the tolerant tester (T2) were highest only for days to anthesis and days to silk (Table 4).The susceptible tester (T3) had the lowest genetic variance estimates for all traits recorded under Striga infestation, except for number of ears per plant.Similarly, the genetic variances estimated for the testcrosses of the resistant tester (T1) were the highest for grain yield and four other traits measured under natural field infestation, while the genetic variances of the tolerant tester (T2) were highest only for ears per plants Striga damage rating at both 8 and 10 WAP and Striga emergence count at 10WAP (Table 4).The susceptible tester (T3) had the lowest genetic variance estimates for all traits recorded under natural Striga infestation.Again, the resistant tester followed by the tolerant tester had the highest genetic variances for most of the traits recorded under non-infested conditions (Table 5).

| Testcrosses means under artificial, natural Striga infestation and non-infested conditions
Yield reduction under Striga infestation relative to that under noninfested condition was 86% for the susceptible check (8338-1) and 53% for the tolerant check (9022-13) (Table S2).The average yield losses due to Striga damage was 15% for testcrosses of the resistant tester (T1), 25% for those of the tolerant tester (T2) and 60% for those of the susceptible tester (T3).Yield increases over the tolerant reference check (9022-23) varied from 17% to 205% for 29 lines crossed to the resistant tester (T1), from 14% to 165% for the 30 lines crossed to the tolerant tester (T2) and from 17% to 106% for only 11 lines crossed to the susceptible tester (Table S2).Among all the testcrosses, 35 produced significantly higher grain yields than the tolerant reference check, whereas 74 had significantly higher grain yields than the susceptible reference check (8338-1).Thirteen of the top 15 testcrosses involving the tolerant and resistant testers were crossed to common inbred lines (L02, L03, L04, L05, L06, L07, L08, L09, L20, L22, L23, L25 and L26).Also, 11 of these parental lines were represented in the top 15 testcrosses involving the susceptible tester (Table S2).On the average, testcrosses of tester T1 produced more grain yield, had shorter anthesis/silking interval, taller plants, better ear aspect and damage scores and supported fewer emerged parasites than those of testers T2 and T3 (Table 6).Furthermore, almost all the T A B L E 4 Genetic variance between testcrosses obtained for the testcrosses of each tester (T1, T2 and T 3) under artificial and natural Striga infestation evaluated at two sites in 2107 and 2018   7).Among all the testcrosses evaluated under natural field infestation, 55 produced significantly higher grain yields than the tolerance reference check (9022-23), whereas 61 had significantly higher grain yields than the susceptible reference check (8338-1).Furthermore, almost all the top 15 testcrosses involving each of the three testers exhibited less Striga damage symptoms and supported fewer emerged Striga plants under natural field infestation (Table S3).
On the contrary, the highest mean grain yield under non-infested condition was observed in testcrosses involving the susceptible tester followed by testcrosses of the lines with the tolerant tester (Table 8).

| Relative ranking of inbred lines across testers and environments under Striga infestation
Spearman's rank correlation between testcrosses involving pairs of testers were significant (P < .001)and varied from .61 to .70 under Striga infestation.Additionally, testcrosses of each of the three testers showed significant concordance of ranks for grain yield and all other traits measured under Striga infestation across environments (Table 9).Testcrosses of T1 showed better concordance coefficients for grain yield and six other Striga-resistance associated traits in comparison to those involving T2 and T3.Testcrosses involving T3 showed the smallest concordance coefficients for grain yield and other traits recorded across environments.Similar consistency of ranks were observed for testcrosses of each tester across environments under natural field infestation.

| Combining ability estimates for testers
Under Striga infestation, the resistant tester had significant and positive GCA effects for grain yield, plant height and ears per plant but negative GCA effects for the remaining traits (Table 10).The tolerant tester also had significant positive GCA effects for grain yield but negative GCA effects for Striga damage rating (Table 10).In   contrast, the susceptible tester had a significant and negative GCA effect for grain yield, plant height and ears per plant but positive GCA effects for Striga damage rating, Striga emergence count and ear aspect (Table 10).Under natural field infestation, the resistant tester again had significant and positive GCA effects for grain yield, but had significant negative GCA effects for days to silking, days to anthesis, anthesis/silking interval, and Striga emergence counts at both 8 and 10WAP.On the other hand, the tolerant tester had significant negative GCA effects for grain yield, but significant and positive GCA effects for days to silking and days to anthesis.The susceptible tester had a significant and positive GCA effect for anthesis/silking interval, Striga emergence counts and a negative GCA effect for plant height (Table 10).Under non-infested condition, the resistant tester had significant negative GCA effects for grain yield and five other traits, while the susceptible tester had significant positive GCA effects for grain yield and three other traits but negative GCA effects for ear aspect and ear height.The GCA effect of the tolerant tester for grain yield under non-infested condition was not significant.Under Striga infestation (Table 11), only two lines crossed to tester T1, three lines crossed to tester T2 and a line crossed to tester T3 had significant positive SCA effects for grain yield.Under natural field infestation, three lines crossed to tester T1 and two lines crossed to tester T3 had significant positive SCA effect for grain yield.Three lines crossed to T1, two lines each crossed to T2 and T3 had significant positive SCA effects for grain yield (Table 11). T

| DISCUSSION
The present study was conducted to determine the relative value of white inbred testers with varying resistance reactions to the parasite in eliciting genetic differences among new resistant maize inbred lines.
The results of analyses found significant differences among testers, indicating that the testers allowed marked expression of genetic variability in grain yield and other Striga-resistance related traits under artificial and natural Striga infestation.The observed significant GCA effects of both lines and testers for most traits imply that the breeding strategy was effective in accumulating favourable alleles with additive effects on traits measured under Striga infested and non-infested conditions.This present the potential that exists to further enhance resistance to Striga in maize using different breeding methods, consistent with findings in other studies that involved other sets of inbred lines (Akanvou et al., 1997;Kim, 1994;Yallou et al., 2009).The significant SCA effects for grain yield and other traits suggest that, promising Striga resistant hybrids can be developed by crossing lines with complementary heterotic groups to optimise expression of heterosis (Hallauer et al., 2010).The line x environment and tester x environment interactions were significant for most traits recorded The GCA effects of testers were greater than that of lines under both artificial and natural Striga infestation, indicating the presence of favourable alleles that impart high grain yield, reduced Striga damage symptoms and less emerged parasites occurring at higher frequencies in the testers than in the lines.Studies have also shown that genetic gain from selection with inbred testers is mainly associated with an additive gene effect rather than non-additive effects (Horner et al., 1976).Similar results were reported in other studies in maize (Akanvou et al., 1997;Badu-Apraku et al., 2007;Gethi & Smith, 2004;Kim, 1994).Among the three testers, the resistant tester (T1) had desirable GCA effects for grain yield, Striga damage rating, emerged Striga count and other traits under both artificial and natural field infestation, indicating that the tester contains high frequency of resistance alleles with additive effects.This tester can then be regarded as a suitable tester for identifying parental lines to develop superior hybrids that can be targeted for commercialization, consistent with the findings of Hallauer and Carena (2009).The tolerant tester also combined positive GCA effect for grain yield with desirable GCA effects for other traits, but not for emerged Striga count.It appears that the tolerant tester contains favourable resistance alleles occurring at intermediate to higher frequency and may also be used as the second potential tester to screen early generation lines under artificial Striga infestation.As suggested by Keller (1949), more than one tester are recommended for accurate assessment of ranks of lines in crosses and also to obtain better variance estimates among testcrosses.
One of the criteria used to identify a good tester is its ability to elicit greater expression of genetic difference among testcrosses of new inbred lines under evaluation (Guimaraes et al., 2012;Hallauer et al., 2010;Matzinger, 1953;Miotto et al., 2016;Russell, 1961).In the current study, testcrosses involving the resistant tester had the largest genetic variance estimates for grain yield and other Strigarelated traits under both artificial and natural Striga infestation as well as under non-infested conditions, while testcrosses of the susceptible tester had the lowest genetic variance estimates for grain yield and other traits under these growing conditions.These results suggest that the resistant tester may either differ from the resistant lines in the frequency of resistance alleles occurring at many loci or carry different sets of alleles at different loci that contributed to the large genetic variances observed among testcrosses.This finding contradicts the reports in other studies (Hallauer & Lopez-Perez, 1979;Rawlings & Thompson, 1962) T A B L E 1 Analysis of variance for combining ability effects of different traits of the 90 testcrosses evaluated under artificial Striga infestation in Abuja and Mokwa in 2017 and 2018 top 15 testcrosses involving each of the three testers exhibited less Striga damage symptoms and supported fewer emerged Striga plants.Similar average performance were observed for testcrosses of tester TI in comparison to testcrosses of testers T2 and T3 under natural field infestation (Tables

T
A B L E 1 0 General combining ability effects of the three testers evaluated under artificial Striga infestation, natural infestation and non-infested conditions in 2017 and 2018 under artificial infestation possibly due to the differential reaction of the lines and the testers to varying infection severity modulated by changes in climatic conditions and soil properties that prevailed in different years and locations(Menkir et al., 2012;Menkir & Meseka, 2019).The non-significant mean squares for line x environment and tester x environment interactions observed under natural infestation could arise from the fact that both Samaru and Agwanmalamai had similar climatic conditions throughout the period of the experiment.Furthermore, examination of concordance in ranking the Striga resistant lines in crosses with each of the three testers were significant for all the traits measured across environments, suggesting that the resistance or susceptibility reactions of testcrosses were consistent under both artificial and natural infestations.The resistant tester followed by the tolerant tester ranked the performance of testcrosses of the resistant lines better across environments than the susceptible tester, which is important in identifying superior single-crosses as female testers to develop three-way cross hybrids for further testing and direct use.In addition, the resistant and tolerant testers identified 13 common inbred lines that formed the top 15 testcrosses under both artificial and natural infestation and 12 common lines in the top 15 testcrosses under non-infested conditions.It is interesting to note that at least nine of the common lines identified by the resistant and tolerant testers were also parents in the top 15 testcrosses involving the susceptible tester.Such resistant maize inbred lines may combine well with other unrelated maize inbred lines for resistance to S. hermonthica and can then serve as excellent sources of resistance to Striga for use in other resistance breeding programmes for tropical lowlands infested with the parasite.
who considered a low-performing tester with a low frequency of favourable alleles at important loci being more effective in differentiating the potential value of inbred lines.It seems that the observed intermediate and low variance estimates for testcrosses of the tolerant and susceptible testers may arise from the occurrence of resistance alleles at intermediate and low frequencies, respectively.The large genetic variance estimates obtained for testcrosses of the resistant and tolerant testers coupled with the identification of the largest number of testcrosses out-yielding the tolerant check by more than 15% suggest that these testers have the potential to effectively discriminate among the resistant maize inbred lines.This is consistent with studies ofFato et al. (2012), who found that the resistant tester was superior to the susceptible tester in identifying maize inbred lines with high levels of resistance to downy mildew and high yield potential in hybrids.
Minimum, maximum and mean values of agronomic traits for the testcrosses of three testers evaluated under artificial Striga infested conditions across four environments Minimum, maximum and mean values of agronomic traits for the testcrosses of three testers evaluated under natural field infestation across four environments T A B L E 5 Genetic variance between testcrosses obtained for the testcrosses of each tester (T1, T2 and tester 3) under Striga non-infested condition evaluated at two sites in 2107 and 2018 maximum and mean values of agronomic traits for the testcrosses of three testers evaluated under Striga non-infested conditions across four environments Kendall's coefficient of concordance of ranking of the 3 testers across the 30 inbred lines for each trait under artificial Striga infestation across four environments A B L E 1 1 Specific combining ability effects of grain yield for the 90 testcrosses evaluated under artificial and natural Striga infestation and non-infested conditions across four environments in 2017 and 2018