HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (10) Citing Articles via Google Scholar Google Scholar Articles by Haussmann, B.I.G. Articles by Geiger, H.H. Search for Related Content PubMed Articles by Haussmann, B.I.G. Articles by Geiger, H.H. Agricola Articles by Haussmann, B.I.G. Articles by Geiger, H.H.

CROP BREEDING, GENETICS & CYTOLOGY

Yield and Yield Stability of Four Population Types of Grain Sorghum in a Semi-Arid Area of Kenya

^a Institute of Plant Breeding, Seed Science, and Population Genetics, Univ. of Hohenheim, 70593 Stuttgart, Germany
^b ICRISAT, P.O. Box 776, Bulawayo, Zimbabwe
^c Dep. of Crop Science, University of Nairobi, P.O. Box 30197, Nairobi, Kenya
^d Agricultural Research Organization, The Volcani Center, P.O.B. 6, Bet Dagan, Israel
^e Südwestsaat GbR, Benshurst 2, 77839 Lichtenau, Germany

geigerhh{at}uni-hohenheim.de

	ABSTRACT

TOP NOTES ABSTRACT INTRODUCTION Materials and methods Results Discussion Conclusions REFERENCES

 
Sorghum [Sorghum bicolor (L.) Moench] is widely grown in semi-arid tropics where local farmers depend on the adaptability of their rainfed crops to unpredictable drought and other stress factors. To investigate the effects of heterozygosity and heterogeneity on the adaptability of grain sorghum, two sets of material, each containing 12 parent lines, six single-cross hybrids, six two-component blends of parent lines, and six two-component hybrid blends were grown in eight macro-environments in the semi-arid Makueni District of Kenya, during 1991 through 1993. Environmental means for grain yield ranged from 584 to 47 g m^-2. In all environments, hybrids outyielded their parent lines, with a mean relative hybrid superiority of 54%. Blending effects were non-significant. Combined analyses of variance were computed with logarithmically transformed data. Entry x environment interaction effects were more important than genetic effects. Lines in pure stand contributed most to the total entry x environment interaction variance. Wide ranges were found within all four groups for stability parameters derived from regression analysis. On average, hybrids in pure stand had most favorable values. Pattern analysis (classification and ordination techniques) was applied to the environment-standardized matrix of entry means from the individual environments. A one-way classification clearly distinguished homozygous from heterozygous entries. Heterogeneous entries were not consistently grouped together. Performance plots for different entry groups showed various patterns of adaptation and illustrated the superiority of heterozygous entries. The biplot from ordination underlined the importance of entry x type-of-drought-stress interaction. Principal Components 1 and 2 were highly correlated with entries' mean yield and regression coefficient, respectively. Breeding heterozygous cultivars could contribute to increased grain yields and improved yielding stability of sorghum in the target area of Kenya.

	INTRODUCTION

TOP NOTES ABSTRACT INTRODUCTION Materials and methods Results Discussion Conclusions REFERENCES

 
SORGHUM IS AN IMPORTANT COMPONENT in traditional farming systems in the semi-arid tropics of Africa and Asia, with mean grain yields of 863 and 1157 kg ha^-1, respectively (FAO, 1999). Low soil fertility, poor stand establishment, and highly variable drought stress are major production constraints in these areas. The local farmers usually do not have access to irrigation facilities or fertilizer stocks and rely totally on the stress resistance and yield stability of their rainfed crop cultivars.

Yielding stability in unpredictably variable environments can be due to individual and/or populational buffering (Allard and Bradshaw, 1964). Initially, the two mechanisms were defined by Lerner (1954) as developmental and genetic homeostasis. Individual buffering may be favored by heterozygosity, and populational buffering by heterogeneity in as much as the different genotypes present in the population are specifically adapted to different environmental conditions (Bradshaw, 1965; Reich and Atkins, 1970; Schnell and Becker, 1986). A concomitant question plant breeders are facing is, what type of genetic structure (i.e., degree of heterozygosity and heterogeneity) is most desirable—or necessary—to obtain improved, stable cultivars. Sorghum, because of the availability of nuclear and cytoplasmic-genic male sterility, offers a wide range of possible genetic structures to the breeder, including homozygous lines, homogeneous or heterogeneous hybrids, as well as homo- or heterozygous, heterogeneous populations or synthetic cultivars.

Estimates of average midparent heterosis for grain yield in sorghum range from 13 to 88%. Lower values were usually obtained with crosses of adapted parent lines (Kambal and Webster, 1966; Kirby and Atkins, 1968; Reich and Atkins, 1970; Liang et al., 1972; Patanothai and Atkins, 1974), while high estimates were most often reported from studies which involved exotic germplasm (Niehaus and Pickett, 1966; Jowett, 1972; Laosuwan and Atkins, 1977), or which were conducted under environmental stress conditions (Jowett, 1972; Kapran et al., 1997).

Five studies were found to deal with genotype mixtures in sorghum. In a trial over 5 yr at one moderate to high-yielding location in Kansas, Ross (1965) obtained 1% superiority of hybrid blends over the mean of their pure stands. The hybrids involved in this study differed in days to anthesis by maximally 11 d. Reich and Atkins (1970) reported a mean blending effect for grain yield of 2%, averaged over homozygous and heterozygous entries and nine moderate to high-yielding environments in Iowa. Results from an experiment grown at four locations in Ghana with environmental means for grain yield between 254 and 2869 kg ha^-1 indicated a slightly larger, but still non-significant yield advantage of two- to five-component blends of local cultivars over the individual pure stands (Mercer-Quarshie, 1979). Marshall and Allard (1974) who studied the performance of lines in pure stand and two-, four-, six-, twelve-, twenty-, fifty-, and eighty-component mixtures in nine non-stress site–season combinations in California stated that favorable blending effects were seldom of sufficient magnitude to offset the burden imposed on heterogeneous populations by the lower yielding genotypes. Yet, Bebawi and Abdelaziz (1983) reported a 25% superiority of two-component blends of cultivars differing largely in their maturity dates over the mean of their pure stands in a study conducted at one location over 2 yr with differing drought stress intensity in Sudan.

Regarding stability of grain yield, Reich and Atkins (1970) reported hybrid blends to be the most productive and stable population type. Jowett (1972) and Patanothai and Atkins (1974) compared heterogeneous three-way crosses with homogeneous single crosses at eight and nine environments in Uganda and Iowa, respectively. Considerable variation was found among the individual hybrids for stability parameters, suggesting that stability of performance may be attainable with both hybrid types. Mercer-Quarshie (1979) reported a trend of increasing stability with increasing complexity of the mixtures in Ghana.

Overall, only limited knowledge is available about the effects of heterozygosity, heterogeneity, and their interaction on the performance of sorghum grown under unpredictable stress conditions. Our experiment was, therefore, designed to examine yield level and stability, and patterns of genotype x drought stress interactions in a genetically balanced system of sorghum entries differing in heterozygosity, heterogeneity, or both in a semi-arid area of Kenya.

	Materials and methods

TOP NOTES ABSTRACT INTRODUCTION Materials and methods Results Discussion Conclusions REFERENCES

 
The basic material consisted of 12 unrelated single-cross hybrids and their respective 24 homozygous parent lines (>F₉ inbreeding generation). These genotypes represented a random sample of actual sorghum breeding germplasm from SADC/ICRISAT (Southern African Development Com-munity/ International Crops Research Institute for the Semi-Arid Tropics) in Zimbabwe. To reduce the size of individual experiments and herewith the experimental error due to natural heterogeneity of the experimental fields, the materials were randomly divided into two sets (Table 1) . Within each set, the following four groups of entry with alternate types of genetic structure were formed, analogous to the experiments of Reich and Atkins (1970) with sorghum and Schnell and Becker (1986) with maize (Zea mays L.):

2. Homogeneous entries of heterozygous plants: six single-cross hybrids among the above 12 parent lines (H1, ..., H6);

3. Heterogeneous entries of homozygous plants: six two-component blends of the parent lines according to the parentage of the hybrids (Lm1, ..., Lm6, corresponding to blends of the lines B1+R1, B2+R2, ..., B6+R6, respectively);

4. Heterogeneous entries of heterozygous plants: six two-component blends of hybrids such, that each hybrid was included in two mixtures (Hm12, Hm34, Hm56, Hm14, Hm25, Hm36, corresponding to blends of the hybrids H1+H2, H3+H4, ..., H3+H6, respectively).

Within each set, the arrangement is genetically balanced in that all four groups of entry share the same set of nuclear genes in equal frequencies. The two sets were each planted together with six standard cultivars (Table 1) in 6 by 6 triple lattice designs side by side in the same field. The standards included the two local cultivars Makueni and Seredo which can be considered improved open-pollinated cultivars with unknown degree of heterozygosity and heterogeneity. The choice of the other standards was determined by seed availability.

The experiment was conducted during 1991 to 1993 in eight macro-environments (site–season combinations) at the semi-arid locations Kibwezi (latitude 2°17'S; longitude 38°01'E; altitude 800 m) and Kiboko (latitude 2°12'S; longitude 37°43'E; altitude 915 m; both Makueni District, Kenya; Table 2) . The long-term average rainfall at Kibwezi and Kiboko amounts to 368 and 320 L m^-2 in the Short Rains and to 239 and 195 L m^-2 in the Long Rains, respectively (Michieka and Van der Pouw, 1977; Ekirapa and Muya, 1991). The annual rainfall at Kibwezi fluctuated from 67 to 1964 L m^-2 in the years from 1919 to 1974.

The computer program PLABSTAT (Utz, 1993) was used for initial statistical analyses. Entries and environments were considered as factors with random effects while the effects of heterozygosity and heterogeneity were regarded as fixed. In a first step, data of each set were analyzed according to the lattice design. Drought intensities were computed relative to the mean of the two highest yielding environments as described by Fischer and Maurer (1978). Repeatabilities (= heritabilities on plot basis) were calculated as outlined by Fehr (1987). Heterogeneity of error variances was tested with Bartlett's Test (Snedecor and Cochran, 1980). Combined analyses of variance across environments were computed with logarithmically transformed data {y = ln[(x/10)+1]} since genetic and environmental effects were related in a multiplicative manner, as indicated by Tukey's (1949) test for non-additivity. Numerical stability parameters were estimated according to Eberhart and Russell (1966) with the logarithmically transformed data. A low regression coefficient, combined with a low deviation mean square and a relatively high grain yield in the low-yielding environments was considered advantageous. Coefficients of phenotypic and genotypic correlation were calculated as described by Mode and Robinson (1959).

Pattern analysis by the software package GEBEI (Watson et al., 1996) was applied to the environment-standardized (Fox and Rosielle, 1982) matrix of the lattice-adjusted entry means from the individual environments. The environment standardization is designed to group environments that are most similar in the way they cause genotypes to be ranked (Fox and Rosielle, 1982). An agglomerative hierarchical procedure with an incremental sum of squares grouping strategy known as Ward's method (Ward, 1963) was employed for the purpose of classification. The squared Euclidean distance was used as dissimilarity measure required in Ward's method. A one-way classification of the entries was chosen, maintaining all eight macro-environments as separate. The number of entry groups was determined considering the percentage of the sums of squares retained in the reduced entry x environment matrix. Performance plots of different entry groups were used to illustrate various adaptation patterns of the entries. The biplot derived from ordination served to assess relations among entries, among environments, and between entries and environments (Kempton, 1984). Entries close to the origin are average in their performance across all environments. Entries or environments that are close to each other tend to be similar. Entries distributed in the increasing direction of an environment vector are positively associated with that environment, while those distributed in the opposite direction are negatively associated. A line drawn perpendicular from a particular entry to an environment vector indicates its relative performance in that environment by the point of intersection, i.e., for the same environment vector, a better genotype would project an intersection point that is farther along in the positive direction of the environment vector. Small angles between two environment vectors indicate strong positive associations, i.e., high similarity between these two environments; 90° angles indicate no association; and angles greater than 90° indicate negative associations.

	Results

TOP NOTES ABSTRACT INTRODUCTION Materials and methods Results Discussion Conclusions REFERENCES

 
Growing conditions during the experimental period were highly variable and were comprised of two non-stress and six stress environments (including preflowering, terminal, and extreme drought). Correspondingly, environmental means for grain yield ranged from 584 to 47 g m^-2 (Table 2). In the following, the two non-stress environments will be named N1 and N2, and the six stress environments D1T, D2P, D3T, D4T, D5T, and D6E, respectively. The D-codes relate to increasing drought intensities and decreasing yield levels; T, P, and E denote presence of terminal, preflowering, and extreme (preflowering and terminal) drought stress, respectively.

Heterozygous entries significantly outyielded homozygous entries in all environments (Fig. 1) , with a mean relative hybrid superiority of 54%, averaged over both levels of heterogeneity. The latter value ranged from 36 to 107% across the individual environments. The highest estimates of hybrid superiority were obtained in the two lowest yielding environments. Differences between blends and pure stands were small and inconsistent across environments for both lines and hybrids. Averaged over the eight environments, two sets of material, and both levels of heterozygosity, the mean blending effect was not significantly different from zero. The same was true for the interactions of heterogeneity with heterozygosity and with environments, and for the respective threefold interaction. Hybrids outyielded the local cultivars Makueni and Seredo by 10 and 14%, respectively.

View larger version (52K): [in this window] [in a new window]   Fig. 1 Means of the four groups of entry and the two local cultivars for grain yield in two non-stress (N1, N2) and six stress environments (D1T, D2P, D3T, D4T, D5T, D6E, with the numbers denoting increasing drought intensity, and T, P, and E terminal, preflowering, and extreme drought stress, respectively) in a semi-arid area of Kenya, averaged over the two sets of material. Vertical bars indicate least significant differences (P = 0.05) among the entry types

The dendrogram for the eight environments derived from pattern analysis clearly separated the extreme stress environment D6E from all other environments in both sets of material (Fig. 2) . The non-stress environments N1 and N2 grouped closely together in both sets of material, whereas the divergence among the drought stress environments D1T, ..., D5T did not fully correspond to the performance level for grain yield or the type of drought stress observed in these environments.

View larger version (10K): [in this window] [in a new window]   Fig. 2 Hierarchical classification of the eight environments obtained from the data of a) Set 1, and b) Set 2. Environment abbreviations as in Fig. 1

View larger version (37K): [in this window] [in a new window]   Fig. 3 Dendrograms showing hierarchical classification of 36 entries in a) Set 1 and b) Set 2, based on environment-standardized grain yield data from eight Kenyan macro-environments. Abbreviations of genotypes as in Table 1

View larger version (25K): [in this window] [in a new window]   Fig. 4 Performance plots of entry groups a) G1, ..., G9 in Set 1 and b) G1', ..., G8' in Set 2. For group compositions see Fig. 3

The corresponding biplots (Fig. 5) showed that there was no correspondence between the discrimination among entries in the environment with solely preflowering drought versus the environment with extreme drought stress in both sets of material (i.e., 90° angle between the two respective environmental vectors). The angles between vectors of the environment with preflowering drought and those with terminal drought were larger than angles between non-stress and the various stress environments, indicating the presence of genotype x type-of-drought-stress interaction. The biplots from both sets of material clearly illustrate the better adaptation of heterozygous entries and local cultivars to the testing environments. Perpendicular lines drawn from homozygous entries meet most environmental vectors at negative values, reflecting the generally poor performance.

View larger version (19K): [in this window] [in a new window]   Fig. 5 Biplots for Principal Components 1 and 2 obtained from the ordination of environment-standardized grain yield data of a) Set 1 and b) Set 2. Environments are characterized by vectors drawn from the origin. For environmental abbreviations see Fig. 1

	Discussion

TOP NOTES ABSTRACT INTRODUCTION Materials and methods Results Discussion Conclusions REFERENCES

 
The target environment of the present study is characterized by extreme variability of environmental conditions within a relatively small area. This was reflected by the huge range of environmental means for grain yield (47–584 g m^-2) obtained during the experimental period of only 2 yr and by the low coefficients of correlation among certain environments. Detailed long-term agroclimatic studies, i.e., with estimates of frequencies and intensities of post planting, preflowering, and terminal drought stress in the area are not available. However, on the basis of rainfall probability estimates, Ekirapa and Muya (1991) stated that farmers in the Kibwezi area suffer from poor crops or crop failures (total rain < 1/2 total potential evaporation) in four out of 10 Short Rains and in six out of 10 Long Rains, while a good crop (total rainfall > 2/3 total potential evaporation) may be expected in four out of 10 Short Rains and only in two out of 10 Long Rains. The probability of crop failures is even higher at Kiboko because of the lower average amount of rain. The impact of crop failures is an important aspect when defining adaptation. Though no socio-economic investigation quantifying the risk aversion of people in the Makueni district of Kenya is available, a high risk aversion of the prevailing subsistence farmers can be assumed. Therefore, cultivars are needed with specific adaptation to extreme stress conditions and broad adaptation to all other, unpredictable types of environments occurring in the area.

Landraces can be considered as the outcome of an evolutionary approach to survival and performance under arid and semi-arid conditions (Ceccarelli, 1994). Thus, their genetic structure must bear some advantage or at least, cannot be "accidental" (Ceccarelli et al., 1991). Hayward and Breese (1993) described the genetic structure of inbred cereal landraces as "stable mixture of coadapted genotypes having different and complementary requirements for resources such as nutrients and light, and thus escaping competitive elimination". A certain degree of heterozygosity may be expected as well, depending on the individual outcrossing rates. In sorghum, outcrossing was rated at 5 to 15% in cultivated sorghum and up to 30% or more in wild relatives (Doggett, 1988).

In the present study, heterozygosity turned out to be an important prerequisite of high grain yield in sorghum grown under drought stress in the Kenyan target area. The obtained mean relative superiority of hybrids over lines for grain yield lies above the average value given for sorghum in the literature. This may be explained by a lack of adaptation of some of the female lines which resulted in a lower midparent performance but was overcome in the hybrids by the genetic contribution of the male parent. The comparatively good performance of the local cultivars Makueni and Seredo and their co-classification with hybrids indeed indicate that the "locals" were better adapted to drought stress than most parent lines. However, the higher yielding ability of the locals could as well have resulted from a certain degree of heterozygosity. In this context, it would have been useful to include a treatment of hybrid blends that had been taken to F₂ or F₃ to obtain checks that were more relevant to the local standards.

The phenomenon of heterosis is particularly important in outcrossing species [e.g., 92–252% in maize (Schnell and Becker, 1986; Melchinger and Gumber, 1998), 81–207 % in rye (Secale cereale L.) (Geiger and Miedaner, 1997), and least important in self-pollinating species [e.g., on average 9% in wheat (Triticum aestivum L.) (Melchinger and Gumber, 1998)]. Partially allogamous species were found to range between the two extremes, as shown for sorghum in this study, or observed for spring faba bean (Vicia faba L.) by Stelling et al. (1994). The larger relative superiority of heterozygotes over homozygotes in outcrossing species appears to be mainly due to their higher degree of dominance (Geiger, 1988) and the concomitant higher mutation load which becomes apparent during the process of inbreeding (Gallais, 1988).

The heterozygous entries were, on average, slightly more stable than the homozygous ones, but the wide ranges found within both lines and hybrids for stability parameters indicate that it should be possible to select stable entries at both levels of heterozygosity. These results agree with the findings of Reich and Atkins (1970), Majisu and Doggett (1972), and Patanothai and Atkins (1974). The effect of heterozygosity on yield stability varies among crop species, depending on their reproductive system (Léon, 1994). In outcrossing species like maize (Schnell and Becker, 1986) and rye (Wahle and Geiger, 1978), heterozygosity has a strong positive effect on yield stability. In partially allogamous species like oilseed rape (Brassica napus L.) (Léon, 1991), faba beans (Stelling et al., 1994), and sorghum (Reich and Atkins, 1970), heterozygosity may improve yield stability, but to a lower extent than in purely outcrossing species, while in inbreeders no benefit from heterozygosity per se on yield stability has been found so far (Léon, 1994).

Blending effects were small, inconsistent, and non-significant for grain yield of both lines and hybrids. Possible reasons for the general lack of blending effects in this study may be that the tested genotypes were not selected for high performance in mixtures, and that they did not differ extremely in developmental or morphological traits. Apart from limited stemborer [Chilo partellus (Swinhoe), Sesamia calamistis Hmps.], shootfly [Atherigona soccata (Rondani)], charcoal rot [Macrophomina phaseolina (Tassi) Goid], or grain mold (Fusarium and Curvularia species) infestation in single environments, no marked diseases or pests were encountered. Therefore, potential advantages of genetic heterogeneity, caused by lower intergenotypic competition, and/or reduced spread of pests and diseases could not materialize. Reviewing results from several crop species, Allard and Hansche (1964) indicated that mean yields of mixtures hold an average advantage in the order of 3 to 5% over the means of their components. Later, Clay and Allard (1969) noted that in only 10% of all reported cases in barley (Hordeum vulgare L.), the superiority of mixtures was statistically significant.

Blending reduced estimated entry x environment interaction variances and deviation mean squares at the homozygous but not at the heterozygous level. Possibly, effective populational buffering requires more diverse hybrids or more complex hybrid mixtures than those evaluated in this study. Similar findings have been obtained for maize by Schnell and Becker (1986). In contrast, Reich and Atkins (1970) and Stelling et al. (1994) found the increased yield stability of hybrid blends in sorghum and faba beans, respectively, to result additively from the effects of heterozygosity and heterogeneity. Yield stability of oilseed rape was reported to profit more from heterogeneity than from heterozygosity (Léon, 1991). However, as the studies mentioned here were not conducted under drought stress, they may not be directly comparable to our results. In barley, some individual lines were shown to be as stable as local heterogeneous populations in Mediterranean, semi-arid climates, at least in the short term (Ceccarelli, 1994).

The biplot visualized large entry x environment interactions for grain yield under preflowering versus terminal and/or extreme drought stress, and extreme versus moderate or non-stress conditions. These relations were also reflected in the phenotypic correlations among the environments. Similarly, Walulu et al. (1994) reported sensitivity to terminal drought of sorghum genotypes with resistance to preflowering stress and vice versa, and Ceccarelli (1994), Ud-Din et al. (1992), and Virk and Mangat (1991) indicated low genetic correlations between grain yield under extreme stress and non-stress conditions for barley, wheat, and pearl millet [Pennisetum americanum (L.) Leeke].

Genetic variation for days to anthesis affected the adaptation to the different drought stress environments in the present study. The best adapted entry groups flowered relatively early, at about 62 to 63 d after planting. In another approach using the same data set, a multiple regression of grain yields under stress on yield potential and days to 50% anthesis was computed for lines and hybrids in pure stands for each stress environment (Haussmann et al., 1998). Differences in yield potential contributed to grain yield differences in all stress environments. Early anthesis was more important than yield potential for specific adaptation to extreme drought. A slight advantage of early anthesis was indicated for environments with terminal drought, while in D2P, characterized by preflowering stress followed by rains, a late onset of anthesis was slightly beneficial.

The present study was not designed to estimate heritabilities for yielding performance at different kinds and intensities of drought stress. However, the repeatabilities given in Table 3 indicate an upper limit on the genetic variability under drought versus non-stress conditions. The repeatabilities for grain yield were highest in the two non-stress environments, but lowest in the third highest yielding environment D1T. These results suggest lower heritabilities under stress relative to non-stress conditions, but do not imply a linear decline of heritabilities with increasing stress intensities. Similarly, heritability estimates were found to be slightly lower under stress conditions for grain yield in wheat (Ud-Din et al., 1992) and pearl millet (Bidinger et al., 1994), and independent from the stress intensity for grain yield of barley (Ceccarelli et al., 1987; Ceccarelli, 1994).

Large genotype x environment interaction effects within a spatially limited area point to the necessity of evaluating breeding materials under a wide range of environmental conditions to achieve broad adaptation. In weighing selection criteria when breeding in semi-arid climates, particular emphasis should be given to specific adaptation to extreme drought stress, since achieving a reliable minimum grain yield is more important to the risk-averse subsistence farmers than high yields in the rare favorable seasons. Given this criterion, the entry groups G7 of Set 1 and G6' of Set 2 may be considered as best adapted to the target area, as they were among the highest yielding groups in the lowest yielding environments, while maintaining an above-average yield in all other environments as well.

	Conclusions

TOP NOTES ABSTRACT INTRODUCTION Materials and methods Results Discussion Conclusions REFERENCES

 
The great hybrid superiority observed in the present study makes hybrid breeding of sorghum for the target semi-arid area in Kenya a promising approach. Our results are supported by the superiority of hybrids over local cultivars in East and West African environments found earlier (Doggett and Jowett, 1966; Kapran et al., 1997). However, one should be aware that hybrid production and successful marketing requires skilled labor, an effective seed industry, a good infrastructure, and a sufficient income for the farmers to be able to afford the costly hybrid seed. These prepositions are not ubiquitous. Instead of hybrids, other types of cultivars could be produced which capitalize on heterozygosity, e.g., synthetics built up from components with superior combining ability and high outcrossing rates. Although synthetic cultivars may not have the same vigor as the hybrid blends evaluated in the present study, they have several advantages. A synthetic cultivar can be regrown for a few seasons without serious changes in its genetic composition, which is convenient for the small-scale farmers. If components are selected for specific adaptation to different types of drought stress, synthetic cultivars can also make use of populational buffering for increased yielding stability.Alagarswamy Chandra 1998

	ACKNOWLEDGMENTS

 
The authors thank the following persons for their contribution to this investigation: the Director of the University of Nairobi Dryland Research Station at Kibwezi; Chris Mukindia and Eli Barak (Kibwezi Irrigation Project); the Directors of the Kenya Agricultural Research Institute (KARI) at Kiboko and Marigat; Mr. Mutinda and Mr. Haggai (KARI); Dr. S.Z. Mukuru and Mr. Kibuka (ICRISAT-Kenya); Anthony Mutie and his family; Daniel Kalua, Vernand Ziro, and all other Kenyans who carefully assisted in the field work. We also like to thank Prof. Dr. H.F. Utz, Dr. G. Seitz, and Dr. H.G. Welz (University of Hohenheim) and Dr. S. Chandra (ICRISAT-India) for most valuable discussions of the data and statistical analyses. Finally, we highly appreciate the generous financial support by the German Israel Agricultural Research Agreement for Developing Countries (GIARA).

	NOTES

TOP NOTES ABSTRACT INTRODUCTION Materials and methods Results Discussion Conclusions REFERENCES

 
ICRISAT J. Paper No. JA 2203.

Received for publication January 29, 1999.

	REFERENCES

TOP NOTES ABSTRACT INTRODUCTION Materials and methods Results Discussion Conclusions REFERENCES

 

• Alagarswamy G., Chandra S. Pattern analysis of international sorghum multienvironment trials for grain yield adaptation. Theor. Appl. Genet. 1998;96:397-405.[ISI]
• Allard R.W., Bradshaw A.D. Implications of genotype-environmental interactions in applied plant breeding. Crop Sci. 1964;4:503-507.[Free Full Text]
• Allard R.W., Hansche P.E. Some parameters of population variability and their implications in plant breeding. Adv. Agron. 1964;16:281-325.
• Bebawi F.Z., Abdelaziz A.H. Grain sorghum responses to pure stands and mixtures under irrigation. Trop. Agric. (Trinidad) 1983;60:262-264.
• Bidinger F.R., Weltzien E., Mahalakshmi R.V., Singh S.D., Rao K.P. Evaluation of landrace topcross hybrids of pearl millet for arid zone environments. Euphytica 1994;76:215-226.
• Bradshaw A.D. Evolutionary significance of phenotypic plasticity in plants. Adv. Genetics 1965;13:115-155.
• Ceccarelli S. Specific adaptation and breeding for marginal conditions. Euphytica 1994;77:205-219.
• Ceccarelli S., Acevedo E., Grando S. Breeding for yield stability in unpredictable environments: Single traits, interaction between traits, and architecture of genotypes. Euphytica 1991;56:169-185.
• Ceccarelli S., Nachit M.M., Ferrara G.O., Mekui M.S., Tahir M., von Leur J., Srivastava J.P. Breeding strategies for improving cereal yield and stability under drought. In: Srivastava J.P., et al. , ed. Drought tolerance in winter cereals. 1987. ICARDA. Chichester, England: John Wiley and Sons Ltd, 1987:101-114.
• Clay R.E., Allard R.W. A comparison of the performance of homogeneous and heterogeneous barley populations. Crop Sci. 1969;9:407-412.
• Cochran W.G., Cox G.M. Experimental designs, 2nd ed London: John Wiley & Sons, Inc, 1957.
• Doggett, H. 1988. Sorghum. 2nd edition. Tropical Agriculture Series, Longman Scientific & Technical, Essex, England.
• Doggett H., Jowett D. Yield of maize, sorghum varieties and sorghum hybrids in the East African Lowlands. J. Agric. Sci. (Cambridge) 1966;67:31-39.
• Eberhart S.A., Russell W.A. Stability parameters for comparing varieties. Crop Sci. 1966;6:36-40.[Abstract/Free Full Text]
• Ekirapa, A.E., and E.M. Muya. 1991. Detailed Soil Survey of Part of the University of Nairobi Dryland Field Station, Kibwezi (Makueni District). Report No. D 57. Kenya Agricultural Research Institute, National Research Laboratories, Nairobi, Kenya.
• FAO, 1999. FAO Quart. Bull. Statis. 12 (1/2):33.
• Fehr W.F. Principles of Cultivar Development. Vol. I: Theory and Technique. New York: Macmillan Publ. Co, 1987.
• Fischer R.A., Maurer R. Drought resistance in spring wheat cultivars. I. Grain yield responses. Aust. J. Agric. Res. 1978;29:897-912.[ISI]
• Fox P.N., Rosielle A.A. Reducing the influence of environmental main effects on pattern analysis of plant breeding environments. Euphytica 1982;31:645-656.[ISI]
• Gallais A. Heterosis: its genetic basis and its utilization in plant breeding. Euphytica 1988;39:95-104.
• Geiger, H.H. 1988. Epistasis and heterosis. p. 395–399. In B.S. Weir, et al. (ed.) Quantitative genetics. Proc. Intl. Conf. on Quantitative Genetics, 2nd, Raleigh, NC. 31 May–5 June 1987. Sinauer Associated Inc., Sunderland, MA.
• Geiger, H.H., and T. Miedaner. 1997. Hybrid rye and heterosis. p. 280–281. In CIMMYT (ed). Book of Abstracts. The Genetics and Exploitation of Heterosis in Crops. An International Symposium. 17–22 Aug. 1997. CIMMYT, Mexico D.F., Mexico.
• Haussmann B.I.G., Obilana A.B., Ayiecho P.O., Blum A., Schipprack W., Geiger H.H. Hybrid performance and relationship among morphological and physiological traits of sorghum [Sorghum bicolor (L.) Moench] under variable drought stress in Kenya. Plant Breeding 1998;117:223-229.
• Hayward M.D., Breese E.L. Population structure and variability. In: Hayward M.D., et al. , ed. Plant breeding: Principles and prospects. London: Chapman & Hall, 1993:16-29.
• Jowett D. Yield stability parameters for sorghum in East Africa. Crop Sci. 1972;12:314-317.[Abstract/Free Full Text]
• Kambal A.E., Webster O.J. Manifestations of hybrid vigor in grain sorghum and the relations among the components of yield, weight per bushel, and height. Crop Sci. 1966;6:513-515.[Abstract/Free Full Text]
• Kapran, I., M. Amadou, M. Abdou, S. Souley, N. Kondo, J.D. Axtell, G. Ejeta, and T. Tyler. 1997. Heterosis and prospects for marketing sorghum hybrids in Niger. p. 258–259. In CIMMYT (ed). Book of Abstracts. The Genetics and Exploitation of Heterosis in Crops. An International Symposium. 17–22 Aug. 1997. CIMMYT, Mexico D.F., Mexico.
• Kempton R.A. The use of biplots in interpreting variety by environment interactions. J. Agric. Sci. (Cambridge) 1984;103:123-135.
• Kirby J.S., Atkins R.E. Heterotic response for vegetative and mature plant characters in grain sorghum Sorghum bicolor (L.) Moench. Crop Sci. 1968;8:335-339.[Abstract/Free Full Text]
• Laosuwan P., Atkins R.E. Estimates of combining ability and heterosis in converted exotic sorghums. Crop Sci. 1977;17:47-50.[Abstract/Free Full Text]
• Léon J. Heterosis and mixing effects in winter oilseed rape. Crop Sci. 1991;31:281-284.[Abstract/Free Full Text]
• Léon, J. 1994. Mating system and the effect of heterogeneity and heterozygosity on phenotypic stability. p. 19–31. In J.W. van Ooijen and J. Jansen (ed.) Biometrics in plant breeding. Proc. meeting of the Eucarpia Section Biometrics in Plant Breeding, 9th. 6–8 July 1994. DLO-Centre for Plant Breeding and Reproduction Research (CPRO-DLO), Wageningen, the Netherlands.
• Lerner M. Genetic homeostasis. London: Oliver and Boyd, 1954.
• Liang G.H., Reddy C.R., Dayton A.D. Heterosis, inbreeding depression, and heritability estimates in a systematic series of grain sorghum genotypes. Crop Sci. 1972;12:409-411.[Abstract/Free Full Text]
• Majisu B.N., Doggett H. The yield stability of sorghum varieties and hybrids in East African environments. East Africa Agric. Forest. J. October 1972;1972:179-192.
• Marshall D.R., Allard R.W. Performance and stability of mixtures of grain sorghum. I. Relationship between level of genetic diversity and performance. Theor. Appl. Genet. 1974;44:145-152.
• Melchinger A.E., Gumber R.K. Overview of heterosis and heterotic groups in agronomic crops. In: Lamkey K.R., Staub J.E., eds. Concepts and breeding of heterosis in crop plants. Madison, WI: CSSA, 1998:29-44 CSSA Spec. Publ. 25..
• Mercer-Quarshie H. Yield of local sorghum (Sorghum vulgare) cultivars and their mixtures in northern Ghana. Trop. Agric. (Trinidad) 1979;56:125-133.
• Michieka D.O., Van der Pouw B.J.A. Soils and vegetation of the Kiboko Range Station. Nairobi, Kenya: Semi-detailed Soil Survey Report No. S3. Ministry of Agriculture - National Agricultural Laboratories, 1977.
• Mode C.J., Robinson H.F. Pleiotropism and the genetic variance and covariance. Biometrics 1959;15:518-537.[ISI]
• Niehaus M.H., Pickett R.C. Heterosis and combining ability in a diallel cross in Sorghum vulgare Pers. Crop Sci. 1966;6:33-36.
• Patanothai A., Atkins R.E. Yield stability of single crosses and three-way hybrids of grain sorghum. Crop Sci. 1974;14:287-290.[Abstract/Free Full Text]
• Reich V.H., Atkins R.E. Yield stability of four population types of grain sorghum, Sorghum bicolor (L.) Moench, in different environments. Crop Sci. 1970;10:511-517.[Abstract/Free Full Text]
• Ross W.M. Yield of grain sorghum (Sorghum vulgare Pers.) hybrids alone and in blends. Crop Sci. 1965;5:593-594.[Free Full Text]
• Schnell F.W., Becker H.C. Yield and yield stability in a balanced system of widely differing population structures in Zea mays L. Plant Breeding 1986;97:30-38.
• Snedecor G.W., Cochran W.G. Statistical methods, 7th ed Ames: The Iowa State University Press, 1980.
• Stelling D., Ebmeyer E., Link W. Yield stability in faba bean, Vicia faba L. 2. Effects of heterozygosity and heterogeneity. Plant Breeding 1994;112:30-39.
• Tukey J.W. One degree of freedom for non-additivity. Biometrics 1949;5:232-242.[ISI]
• Ud-Din N., Larver B.F., Clutter A.C. Genetic analysis and selection for wheat yield in drought stressed and irrigated environments. Euphytica 1992;62:89-96.
• Utz, H.F. 1993. PLABSTAT: A computer program for the statistical analysis of plant breeding experiments. Institute of Plant Breeding, Seed Science, and Population Genetics, University of Hohenheim, Stuttgart, Germany.
• Virk D.S., Mangat B.K. Detection of cross over genotype x environment interactions in pearl millet. Euphytica 1991;52:193-199.
• Wahle G., Geiger H.H. Vergleich der phänotypischen Stabilität von Inzuchtlinien, Einfachhybriden und Populationen bei Winterroggen. Z. Pflanzenzüchtg. 1978;80:211-222.
• Walulu R.S., Rosenow D.T., Wester D.B., Nguyen H.T. Inheritance of the stay green trait in sorghum. Crop Sci. 1994;34:970-972.[Abstract/Free Full Text]
• Ward J.H. Hierarchical grouping to optimize an objective function. J. Am. Stat. Assoc. 1963;58:236-244.[ISI]
• Watson S.H., DeLacy I.H., Podlich D.W., Basford K.E. GEBEI. An analysis package using agglomerative hierarchical classificatory and SVD ordination procedures for genotype x environment data. Brisbane, Australia: The University of Queensland, 1996 Research Report #57, Department of Mathematics.

This article has been cited by other articles:

				M. Li, X. Chen, and J. Meng Intersubgenomic Heterosis in Rapeseed Production with a Partial New-Typed Brassica napus Containing Subgenome Ar from B. rapa and Cc from Brassica carinata Crop Sci., January 24, 2006; 46(1): 234 - 242. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (10) Citing Articles via Google Scholar Google Scholar Articles by Haussmann, B.I.G. Articles by Geiger, H.H. Search for Related Content PubMed Articles by Haussmann, B.I.G. Articles by Geiger, H.H. Agricola Articles by Haussmann, B.I.G. Articles by Geiger, H.H.