Previous PageTable Of ContentsNext Page

Combining ability of CIMMYT’s early maturing maize (Zea mays L.) germplasm under stress and non-stress conditions and identification of testers

Alexander Pswarayi1 and Bindiganavile Vivek2

1 Universitat de Lleida, Department Produccio Vegetal i Ciencia Forestal, Avinguda Rovira Roure, Numero 191, 25198 Lleida, Spain.
CIMMYT, P.O. Box MP 163, Mt. Pleasant, Harare, Zimbabwe; Email


The area sown to early-maturing maize varieties (maize that reaches physiological maturity in about 120 days at 1500 m above sea level, latitude 17.48 S, longitude 31 E (Harare, Zimbabwe), in the main season) across seven southern African countries is at least 4.2 million ha. Although long-season types yield more under favourable conditions, the demand for early-maturing maize is high. Heterotic relationships amongst CIMMYT’s early-maturity germplasm are not well known and early-maturing testers have not been identified. The single crosses CML312/CML442 (heterotic group A) and CML395/CML444 (heterotic group B) currently being used as testers are intermediate and late in maturity, respectively. This slows the identification of early maturing three-way cross hybrids directly from the phase of early generation testing of inbred lines. In an effort to identify earlier maturing replacements for the above testers, 66 testcrosses (generated from a twelve-parent diallel using one OPV and 11 inbred lines) were evaluated at two optimal, one low nitrogen and one droughted environment in Zimbabwe in 2003. The single cross (L7/L8) was identified as a potential new tester for group A because (a) inbred lines L7 and L8 belonged to the heterotic group A; (b) both L7 and L8 had good GCA effects for grain yield; (c) the hybrid, L7/L8 had good yields: 9.8 t/ha (optimal), 3.4 t/ha (low nitrogen) and 2.1 t/ha (drought), an important consideration for selecting single-cross testers.

Media Summary

We report a potential new single-cross tester that could speed up the development of early maturing maize (Zea mays L.) hybrids in southern Africa.

Key Words

Tester, Early Maturity, Combining Ability, Zea mays, Heterotic Groups, CIMMYT


Maize is an important food crop in sub-Saharan Africa providing 50% of the calories in diets in southern Africa, 30% in eastern Africa and 15% in West and Central Africa. Many African countries frequently experience maize shortages and approximately 100 million people are malnourished in this region. Average maize yield is 1.3 t/ha (FAO statistics:

Some of CIMMYT’s efforts are directed towards developing early-maturing maize germplasm (varieties that flower between 55 to 60 days, and mature physiologically at 120 days, after emergence, at Harare, Zimbabwe (1500 meters above sea level, latitude 17.48 S and longitude 31 E), in the main season) (CIMMYT 2000). In Eastern and Southern Africa, early maturing varieties are planted to an estimated total area of 2.7 million ha which translates to 3.5 million tons of grain annually capable of feeding 40 million people per year (average consumption - 87 kg/person) (Pingali 2001). Farmers grow early maturing maize varieties because such varieties:

  • are ideal for off-season plantings in drying riverbeds.
  • provide an early harvest to bridge the “hungry-season” before harvest of a full-season crop. This is especially important in areas where there are two growing seasons.
  • can be used to produce a crop during the secondary, short rains, which enables the plating of a full season maize crop or other crops in the following main season.
  • are ideal for intercropping as they provide less competition for moisture, light, and nutrients than later maturing varieties (CIMMYT 2000).
  • offer flexibility in planting dates, which enables: (i) multiple plantings in a season to spread risk of losing a single crop to drought (ii) late plantings during delayed onset of rainfall (iii) avoidance of known terminal drought periods during the cropping season.

CIMMYT’s early maturing maize breeding program is relatively young, with little information on combining abilities of lines and hence lacks early maturing testers. Single cross testers, CML312/CML442 (group A) and CML395/CML444 (group B), are currently being used for early generation testing of inbred lines. Any good three-way cross that is identified is a potential product as three-way and double-cross hybrids (and not single cross hybrids) are the products commercially sold in most of sub-Saharan Africa. The drawback of using the above testers in an early maturity maize breeding program is that CML312/CML442 is intermediate in maturity and CML395/CML444 is late in maturity making the hybrids selected later in maturity than desired. The objectives of this study were: (1) to determine heterotic relationships amongst 11 inbred lines and one open pollinated variety, (2) to identify early maturing testers.

Material and Methods

Twelve parents (Table 1) were crossed in the winter of 2002 at Mzarabani in a diallel to give 66 crosses excluding reciprocal crosses. Traits for cross of line 12 line 18 was estimated as a missing single cross by the following formula (Eckhardt 1942): (n-1)(Ta+Tb)-2T/n2 –5n+6; where: n is the number of inbred lines in the set, Ta and Tb are totals of performances from different traits including the missing cross, 2T is the grand total from all the crosses for each trait in the set. ZEWA (an early maturing OPV from heterotic group A) and CML 395 (a late maturing inbred line from heterotic group B) were included in the diallel as a standard for heterotic group classification. The germplasm was from diverse sources and of different maturity periods (early, intermediate and late). The 66 testcrosses were evaluated during the summer of 2003 (optimal and low N) and winter of 2003 (drought).

Table 1. Pedigree of parents used in the diallel.

Griffing’s method 4 (Griffing 1956), adapted to a SAS program for diallel analysis, was used to determine the GCA and SCA effects. Only the F1’s, and neither parents nor reciprocals were included in the analysis. Trials of inbred parents were planted close to the hybrid trials to enable calculation of heterosis.

Results and Discussion


Over environments, differences among entries (genotypes) were significant (Ρ<0.01) for: grain yield (GY), anthesis date (AD) (Table 2), plant height (PH), ear position (EPO), husk cover (HC), texture (TEX) and plant height (PH), and significant (Ρ<0.05) for anthesis-silking interval (ASI) and root lodging (RL). GY, PH and EPO were significant at all four sites; AD and ASI were significant at three sites; RL, EPP, HC and TEX were significant at two sites. Significant differences among entries for the different traits enabled identification of single crosses that could potentially be used as testers.

Table 2. Analysis of variance for grain yield (GY) across 4 sites and anthesis date (AD) across 3 sites in Zimbabwe.

Heterotic Relationships

Lines 4 and 6 (Table 1) were classified into group B based on positive SCA effects for GY with ZEWA and a significantly poor SCA effect with CML395. Line 7 was confirmed to be heterotic group A. Positive SCA effects between inbred lines generally indicates that lines are in opposite heterotic groups (Vasal et al. 1992). Lines in the same heterotic group tend to exhibit negative SCA effects when crossed together. Heterotic grouping of lines 2, 3, and 10 were undetermined because the hybrids had negative SCA effects with both testers. Lines 5 and 9 were classified into group AB because their hybrids with both testers had positive SCA effects for GY. Warburton et al. (2002) pointed out that a line could be classified into a new group if it had positive SCA effects with testers from opposite groups. Lines 5 and 9 were further classified into heterotic groups AB1 and AB2 respectively, because the hybrid L5/L9 had positive SCA effect for GY (0.53 t/ha), which indicated that there was sufficient heterotic separation with group AB, justifying the need to categorize them into sub-groups. Classification of line 4 was changed from the previous classification. Testers used in this study, which were different from those used in the previous classification, could have likely caused this change. The changing of groups also showed that heterotic groups are not absolute as was pointed out by Hallauer et al. (1988).


Single cross L7/L8 was identified as a potential tester for group A because it combines the following (Table 3):

  • Classification of the inbreds constituting the hybrid (L7 and L8) to the same heterotic group (group A) as a result of distinct separation by CML395 and ZEWA
  • Good GCA effects for grain yield of inbred lines L7 and L8
  • Stability in yield under diverse environments
  • Lack of inbreeding depression (hybrid L7/L8 showed reasonable intra-group heterosis and positive SCA effects in all environments).
  • Earlier in maturity (65 days to anthesis) compared to the existing A type tester CML312/CML442 (72 days to anthesis)

Hybrid L4/L12 combines most of the above points mentioned for it to be considered as another potential tester. However, it may be less suitable as a tester as L4 and L12 have poorer GCA effects for grain yield compared to L7 and L8. L4/L12 is a top-cross (inbred line L4 crossed to OPV L12). Most practical breeding programs may not find this to have added advantage as this would involve generating an extra cross for little potential gain in yield for use as a female parent. The use of the OPV (L12) as a female parent in the commercial production of top-cross hybrids may have more benefits.

L2/L8 is another candidate that would be worth considering as an early maturing tester. This hybrid does have several attributes mentioned above to qualify as a tester. However, the fact that L2 is recognized as a good general combiner that tends to produce good hybrids with lines from both groups A and B (hence classified as heterotic group AB) may make it less suitable as a tester. This is confirmed by the lack of distinct heterotic separation of L2 with L1 and L12.

Table 3. GCA effects, SCA effects and grain yield of best 12 and worst 12 hybrids under optimal, low nitrogen and drought environments.

Further investigations with newer inbred lines are required to identify testers for group B. Vasal et al. (1992) have also reported difficulty in identifying many good lines in tropical maize germplasm.


Single cross L7/L8 was identified as a potential tester for group A. The heterotic relationships established in this study were: group A - lines 7, 8, 11 and 12; group B - lines 1, 4 and 6; group AB1 - line 5 and group AB2 - line 9. Further investigations are required to identify an early maturing tester for heterotic group B.


CIMMYT-Zimbabwe (2000). CIMMYT-Zimbabwe: 2000 Research Highlights. Harare: Zimbabwe.

Eckhardt RC (1942) Predicting Yields of Missing Single Crosses of Corn. Journal of the American Society of Agronomy 34, 923-932.

FAO Statistics;

Griffing B (1956) Concept Of General And Specific Combining Ability In Relation To Diallel Crossing Systems. Australian Journal Of Biological Science. 9, 463 – 493.

Hallauer AR, Russel WA and Lamkey KR (1988) Corn Breeding. pp 469-565. In G.F. Sprague and J.W. Dudley (ed) Corn and Corn Improvement. ASA, CSSA, and SSSA, Madison, WI.

Pingali PL (2001) CIMMYT 1999-2000 World Maize Facts and Trends. Meeting World Maize Needs: Technological Opportunities and Priorities for the Public Sector. Mexico, D.F.: CIMMYT.

Vasal SK, Srinivasan G, Pandey S, Cordova HS, Han GC and Gonzalez FC (1992) Heterotic Patterns Of Ninety-Two White Tropical CIMMYT Maize Lines. Maydica 37, 259 – 270.

Warburton ML, Xianchun X, Crossa J, Franco J, Melchinger AE, Frich M, Bohn M and Hoisington D (2002) Genetic Characterization of CIMMYT Inbred Maize Lines and Open Pollinated Populations Using Large Scale Fingerprinting Methods. Crop Science 42, 1832-1840.

Previous PageTop Of PageNext Page