 |
| About Bioline | All Journals | Testimonials | Membership | News |
|
||||||
|
||||||
Diallel analysis for reaction to Exserohilum turcicum of maize cultivars and crosses H.F. OJULONG, E. ADIPALA and P.R. RUBAIHAYO Department of Crop Science. Makerere University, P.O Box 7062, Kampala, Uganda (Received 9 September, 1994; accepted 26 November 1995)
ABSTRACT
The effectiveness of a diallel cross in initiating recurrent selection as a breeding procedure for concentrating genes for resistance to Exserohilum turcicum (Pass) Leonard and Suggs was studied in maize single and double cross F1's during the first rains of 1993. Resistance was expressed as reduced percentage leaf area blighted, area under disease progress curve of percentage leaf area blighted (AUDPC-S) and lesion numbers (AUDPC-L), and apparent infection rate (r), indicating presence of rate-reducing resistance. General combining ability effects differed with the genotype, cross and disease indices, resistant cultivars giving the highest negative values, and the single cross giving higher negative values than their corresponding double cross gemstypes. High heritability values were obtained for both types of crosses, signifying the highly heritable nature of this polygenic resistance and the low level of resistance present in the parent cultivars. Key Words: Combining ability, diallel cross, disease progress, northern leaf blight, Zea mays RESUME L'efficiencience d'un croisement diallele en selection recurrante comme procedure d'amelioration en vue de concentrer des genes de resistance a l'Exserohilum turcicum etait etudiee en croisement simple et double de mais de F1's durant la premiere saison pluvieuse en 1993. La resistance etait exprimee en fonction de la reduction du pourcentage de feuilles attaquees, de la courbe de progression de la surface malade, du pourcentage de feuilles attaquees (AUDPC-S) et du nombre de lesions (AUDPC-L), ainsi que du rapport d'infection apparante (r) indiquant la presence d'un rapport de reduction de resistance. Les effets generaux d'habilite combinee etaient differents selon le genotype, le croisement et les incidences de la maladie. Les cultivars resistants donnent des hautes valeurs negatives tandis les croisements simples donnent les valeuL~ ~egatives plus elevees que leurs correspondants genotypes a croisements doubles. Des valeurs elevees d'heritabilite etaient obtenues pour les deux types de croisement, ce qui explique la nature hautement heritable de cette resistance polygenique et le faible niveau de resistance present chez les parents de cultivars. Mots Cles: Habilite combinee, croisement diallele, progression de la maladie, Zea mays INTRODUCTION The most economic control of northern leaf blight (NLB) of maize caused by the fungus Exserohilum turcicum (Pass) Leonard and Suggs, is the use of genetic resistance in the host plants (Smith and Kinsey, 1980). The unexpected outbreak of NLB in Uganda in 1988 stimulated pathologists and plant breeders to identify sources of resistance and improve the resistance of maize cultivars grown in the country. High level of field resistance was identified in Babungo 3 and EV8342-SR, while other cultivars had varying but little resistance (Adipala et al., 1993). In Adipala et al.'s (1993) study, final percentage leaf area blighted, apparent infection rate (r), and area under disease progress curve (AUDPC) were appropriate indices for assessing reactions of maize plants to E. turcicum. A number of workers have used dialtel cross analysis for initiating recurrent selection for disease resistance in maize and other crops (Balkema-Boomstra and Masterbroek, 1993; Barten et al., 1993; Gevers et al., 1994). Dialtel crosses give the largest cross combinations from which selection can be made, and enable the identification of cultivars with good combining abilities especially in uncharacterised germplasm, which can then be used in breeding programmes (Barten et al., 1993; Gevers et al., 1994). It also enables a breeder to predict offspring performance from parental performance. In view of the potential destructiveness of NLB, diallel crosses were made and used to identify cultivars with good combining abilities and potential crosses for breeding programmes. MATERIALS AND METHODS A recurrent selection programme to concentrate resistance genes to E. turcicum in maize in Uganda was initiated during the second rains of 1992 by use of a diallel cross. Ten improved populations described by Adipala et al. (1993) were used in the study. These improved populations have a high degree of homogeneity, because each represents reconstituted full-sib populations selected for diverse agronomic traits, hence, may be called cultivars (Adipala et al., 1993 ). Parental cultivars were used to make single diallel crosses. The F1's made the previous season by crossing the respective cultivars with Babungo 3, were used to make double diallel cross hybrids. In 1993, the resulting crosses were planted in a randomised complete block design with three replicates. The double cross hybrids were planted in a field which had maize the previous season and the single cross planted on the adjacent plot. Three cobs per genotype were planted ear4o-row at a spacing of 0.75 x 0.30 m. Plants were artificially inoculated at growth stages (GS) 6 and 7 (Ritchie and Hansay, 1982) using inoculum prepared from infected leaf tissue collected from the field (Adipala et al., 1993). Ten plants in each row were rated for disease reaction at GS 8 and, thereafter, weekly for a total of five times. Two leaves per plant were assessed, the one immediately above and below the ear leaf, based on their impact on yield (Bowen and Pedersen, 1988). A scale of 0-75 %, based on the portion of the leaf tissue blighted, was used for assessing percentage leaf area blighted. The number of lesions per leaf was determined by averaging the number of lesions on the two leaves. Twenty clearly expressed lesions per line were measured longitudinally (in cm) to give lesion length. Ten ears per line were harvested at physiological maturity, sun-dried, and grain weight determined. Analysis of variances were performed for disease ratings and yield. Lesion counts and percentage leaf area blighted data were used to calculate area under disease progress curve (AUDPC), apparent infection rate (r) and intercept of the linearised logistic model (Y0*) as described by Van der Plank (1963) and Campbell and Madden (1990). AUDPC was standardised by dividing the AUDPC value by the total time duration (t[n]-t[1]) of the epidemic (Fry, 1977), to enable comparisons between the two sets of experiments (Campbell and Madden, 1990). Disease ratings were made the same day for both experiments so that y0* could be used for comparison between the experiments (Campbell and Madden, 1990). General combining ability (GCA) and specific combining abllily (SCA) erieors of the parental genotypes were estimated in the F1 hybrids using Griffing's (1956) method 2 model 1 (includes F1's plus parents with no reciprocal crosses). A computer program, MSTATC (Russel and Freed, Michigan University), menu DIALLEL, was used to estimate combining abilities. The relative importance of parental GCA effects in predicting hybrid performance was evaluated using the variance ratio 2VGCA/(2VGCA+VSCA) (Baker, 1978). Heritability (h^2) estimates were calculated from estimated variance components (Wright, 1985; Barten et al., 1993). RESULTS
The levels of NLB developed and differences in AUDPC of
percentage leaf area blighted (AUDPCS), AUDPC of lesion
numbers (AUDPC-L) and lesion length were significant
(P
TABLE 1. Analysis of variance for the different northern leaf
blight disease variables assessed on the F1's and parents of a
10 x 10 single diallel cross experiment during the first rains
of 1993 at Kabanyolo, Uganda
^a Replication was considered a random effect
TABLE 2. Anaysis of variance for the different northern leaf
blight disease variables assessed on the F1's and parents of a
8 x 8 double diallel cross experiment during the first rains
of 1993
Hybrids were not significantly different from parents,
suggesting that there was no dominance gene action. Hybrids
had scores which were intermediate of the parents. On
average, the hybrids were not significantly different from the
moderately resistant parents. In almost all cases, hybrids
formed from a susceptible recipient parent and a resistant
donor parent were significantly (P
TABLE 4. Average area under disease progress curve for lesion
numbers for hybrids in single and double diallel crosses
during the first rains of 1993
^a Calculated as described by Griffing (1956)
TABLE 5. Mean squares from combining ability analysis of
variance for disease reaction to Exserohilum turcicum
in the F1 generation in single and double diallel maize
crosses grown during the first rains of 1993
General combining ability and Specific combining ability
estimated as described byGriffing (1956), method 2, model 1
correlation between resistance andyield (Ceballos et al.,
1991). Specific combining ability (Table7) followed the
same trend as GCA, bill in almost all cases SCA had
lower negative values compared with the GCA values, probably
indicating that SCA effects were less important.
High heritability variances were recorded for both crosses
(Table 8). The single crosses had higher values for almost all
disease indices, while the reverse was again true for yield.
AUDPC-'S and AUDPC-L had the highest heritability values,
while yield had the lowest.
Yield differences were significant (P<0.05) among
genotypes in both crosses, but "Within parents" was highly
significant (P<0.001) in the single cross only. Yield
differences were significant (P
TABLE 6. Estimates of general combining ability (GCA)a
effects of parental genotypes in the F1 generation of single
and double crosses during the first rains of 1993 for disease
reaction to Exserohilum turcicum
^a Calculated as described by Griffing (1956)
TABLE 7. Estimates of specific combining ability (SCA)^a
effects of parental genotypes in the F1 generation of single
anddouble crosses during the first rains of 1993 for disease
reaction to Exserohilum turcicum
^a Calculated as described by Griffing (1956)
DISCUSSION
The increase in resistance of the susceptible cultivars
following crossing with the resistant parents is an indication
of the very low levels of resistance in the maize cultivars,
which were selected mainly for resistance to the maize streak
virus and for high yield (Adipala et al., 1993;
Ceballos et al., 1991).
The lack of significance of y[o]* and significance of r
values, support the earlier report by Adipala et al.
(1993) that the Uganda cultivars possessed varying levels
of rate reducing resistance to E. turcicum. The results
further show that the resistance is additively inherited, a
further testimony to the polygenic nature of the resistance
(Hughes and Hooker, 1971). Mean squares for GCA were much
greater than those for SCA (Table 5). Lira (1975), when
evaluating a diallel cross of maize inbreds to
Helminthosporium maydis reported similar results and
concluded a predominance ofadditive gene effects for disease
reststance expressed by the genotypes. In addition the results
confirm the highly heritable nature of the E.
turcicurn-maize polygenic resistance pathosystem (Jenkins
et al., 1954; Ceballos et al. 1991). The results
contradict earlier literature that breeding for resistance
reduces yield potential, as this would have resulted in very
low heritability values for yield. This is most likely due to
the enhanced resistance in the double cross which provided
protection under the high disease press ure induced by
artificial inoculation. Similar results were reported by Miles
et al. (1981) and Ceballos et al. (1991). The
ratio 2VGCA/(2VGA+VSCA) was close to unity in both cases
(Table 8), indicating that the hybrid performance could be
predicted to a great extent by GCA effects (Barren et al.,
1993). This could be a result of little variation in
resistance of the double cross parents which was brought about
by crossing with Babungo 3.
The low ratings of crosses involving KWCASR could be due to
its wide genetic base (Adipala et al., 1993), which
possibly contained alleles for resistance to E. turcicum.
Secondly, all the parents except KWCA-SR are foreign
introductions and therefore differ in genetic background.
When these cultivars are crossed with KWCA-SR there is a
combination of genes from diverse gene pools leading to a
bigger gene pool frtm where resistance genes are drawn. Since
polygenic resistance is additively inherited (Hughes and
Hooker. 1971 ). this is likely to result in greater resistance
of the offsprings. KWCA-SR showed good combining abilities
with most of the cultivars, and the offsprings showed high
levels of resistance. Therefore, future crosses should involve
KWCA-SR.
TABLE 8. Estimated variance components for northern leaf
blight and yield in a single and double diallel crosses during
the first rains of 1993
^a True variance component of GCA and
The diallel analysis for combining ability indicated that
genetic variation for disease resistance in the cultivars
studied was associated with highly significant GCA effects.
Resistant cultivars (Babungo 3, EV8342-SR and H99) had the
highest heritability values and can be used in future breeding
programmes as sources of resistance. In the single cross, H99
had the best GCA effects while in the double cross it was the
cross KWCA-SR x Babungo 3. This represents a potential
advantage to breeders, because the cultivars H99 and KWCA-SR
have potential agronomic traits which can be exploited. H99 is
very short (0.5 m) while KWCA-SR is tall (about 3 m) but is
popular with Ugandan farmers. Crossing these two cultivars
resulted in increased resistance and reduction in height for
KWCA-SR (data not shown), since both traits are additively
inherited. SCA effects were also significant but were less
important than the GCA effects indicating that dominance
effects played part in inheritance of this disease although to
a less extent (Barren et al., 1993; Gevers et al.,
1994).
The primary objective of the recurrent selection was to
increase resistance to NLB instead of yield. However, yields
of most lines were comparable. Usually in absence of disease
challenge the susceptible lines do better than the resistant
ones, but under strong disease pressure, as was the case here,
the protection provided by the resistance resulted in higher
yields. These results indicate that the selection for disease
resistance did not affect yield potential, the performance of
the lines being dependent on disease pressure (Miles et
al., 1980; 1981; Ceballos et al., 1991).
All the variables were able to predict hybrid performance,
as shown by the variance ratio which tended to unity, and took
into consideration leaf position in relation to the impact on
yield (Bowen andPedersen, 1988; Barten et al., 1993).
However, based on the GCA effects AUDPC-S was the best index
for disease assessment and takes less time, while lesion
length was the least reliable. Also the disease ratings based
on the two leaves required less time, and takes into account
leaf position in relation to impact on yield.
CONCLUSION
The results of the present study indicate that it is possible
to incorporate resistance to NLB in the Uganda germplasm using
recurrrent selection. It also confirms the highly heritable
nature of the NLB - maize pathosystem. Of the cultivars
tested, KWCA-SR had the best combining ability and could be
used in future crossing programmes.
ACKNOWLEDGEMENT
Funding was provided by the United States Agency for
International Development (USAID)/Uganda Manpower for
Agricultural Development (MFAD). The authors thank Dr. J.J.
Hakiza of Kalengere Highland Crop Research Station, Uganda,
for his useful comments and suggestions.
REFERENCES
Adipala, E., Lipps, P.E. and Madden, L.V. 1993. Reaction
of maize cultivars from Uganda to Exserohilum turcicum.
Phytopathology 83: 217-223.
Baker, R.J. 1978. Issues in diallel analysis. Crop Science
18:533-536.
Balkema-Boomstra, A.G. and Masterbroek, H.D. 1993. Diallel
analysis of powdery mildew caused by Erysiphae graminis
f.s. hordei in spring barley (Hordeum vulgare
L.). Euphytica 65:15 -21.
Barren, J.H.M., Elkind, Y., Scott, J.W.; Vidavski, S. and
Kedar, N. 1993. Diallel analysis over two environments for
blossom-end scar size in tomato. Euphytica
65:229-237.
Bowen, K.L. and Pedersen, W.L. 1988. Effects of northern leaf
blight and detasselling on yields and yield components of corn
inbreds. Plant Disease 72:952-956.
Campbell, C.L. and Madden, L.V. 1990. Introduction to Plant
Disease Epiderniology. John Wiley and Sons, New York.
532pp.
Ceballos, H., Deutsch, J.A. and Gutierrez, H. 1991. Recurrent
selection for resistance to Exserohilum turcicum in
eight subtropical maize populations. Crop Science
31:964-971.
Fry, W.E. 1977. Integrated control of potato late blight
effects of polygenic resistance and techniques of timing
fungicide applications. Phytopathology 68:1650-1655.
Gevers, H.O., Lake, J.K. and Hohls, T. 1994. Diallel cross
analysis of resistance to gray leaf spot in maize. Plant
Disease 78:379-382.
Grilling, B. 1956. A generalized treatment of the use of
diallel crosses in quantitative inheritance. Heredity
10:31-50.
Hughes, G.R. and Hooker, A.L. 1971. Gene action conditioning
resistance to northern leaf blight in maize. Crop Science
11:180-184.
Jenkins, M.T. and Robert, A.L. 1952. Inheritance of resistance
to the leaf blight of maize caused by Helminthosporium
turcicum. Agronomy Journal 46:438-442.
Jenkins, M.T., Robert, A.L. and FindIey, W.R., Jr. 1954.
Recurrent selection as a method of concentrating genes for
resistance to Helminthosporium turcicum leaf blight in
corn.
Agronomy Journal 46:89-94.
Lim, S.M. 1975., Diallel analysis of reaction of eight corn
inbreds to Helmin thosporium maydis race T.
phytopathology 65: 10-15.
Miles, J.W., Dudley, J.W., White, D.G. and Lambert, .R.J.
1980. Improving corn population for grain yield and
resistance to leaf blight and stalk rot. Crop Science
20:247-251.
Miles, J.W., Dudley, J.W., White, D.G. and Lambert, R.J. 1981.
Response to selection for resistance to four diseases in two
corn populations. Crop Science 21:980-983.
Nelson, R.R. 1975. Horizontal resistance in plants: Concepts,
controversies, and applications. In: Proceedings of the
Seminar on Horizontal Resistance to the Blast of Rice.
Galvez, G.E. (Ed.), pp. 1-20. Set. CE-(Cali, Columbia:
CIAT).
Nelson, R.R. 1976. Genetics of horizontal resistance to plant
diseases. Annual Review of Phytopathology
80:309-313.
Ritchie, S.W. and Hansay, J.J. 1982. How a corn plant
develops. Iowa State University Science Technical
Cooperative Extension Special Report 48.
Smith, D.R. and Kinsey, J.G. 1980. Further physiological
specialization in Helminthusporium turcicum. Plant Disease
64:779-781.
Takan, J.P., Adipala, E. and Ogenga-Latigo, M.W. 1994.
Progress and spread of Exserohilum turcicum from
infested maize residue. African Crop Science Journal
2:197-205.
Van dev Plank, J.E. 1963. Plant Diseases: Epidemics and
Control. Academic Press, New York. 349pp.
Wright, A.J. 1985. DiaIlel designs, analysis, and reference
populations. Heredity 54:307-311.
Copyright 1996 The African Crop Science Society
|
| |||||||||
