About Bioline  All Journals  Testimonials  Membership  News  Donations

African Crop Science Journal
African Crop Science Society
ISSN: 1021-9730 EISSN: 2072-6589
Vol. 9, Num. 3, 2001, pp. 463-470

African Crop Science Journal, Vol. 9. No. 3, pp. 463-470



International Institute of Tropical Agriculture (IITA), PMB 520, Ibadan, Nigeria
1Department of Crop Science, University of Calabar, Calabar, Nigeria
2Department of Agronomy, University of Ibadan, Ibadan, Nigeria

Received 11 October, 1999
Accepted 31 March, 2001

Code Number: cs01065


Field studies were conducted to evaluate the segregation of the F3 (early generation) and F6 (late generation) families for seed weight, pod length and days to flowering among cowpea inter-sub-specific crosses. A wide range of segregants were provided in this cross and families were highly significantly different in the three agronomic traits studied. The continuous distributions observed for these traits studied in both generations confirms the quantitative nature of inheritance for these traits. Broad sense heritability estimates ranged from 47.8 to 91.1%. Estimates of genetic advance ranged from high to low and were consistent in both generations for all the traits. The F3 and F6 generations were not significantly different in all the three agronomic traits. Intergeneration correlations ranging from 0.35 to 0.49 also revealed strong associations between traits measured in the two generations. A no significant drop was observed between F3 mean and the corresponding F6 mean. This suggests the existence of a good measure of additive and possibly of additive x additive components of variance (which alone are fixable through subsequent inbreeding) although some amount of dominance and duplicate epistasis (which are non-fixable) may also be operative. The results of this study indicate that selection in early generations for superior types is feasible.

Key Words: Agronomic traits, early generation, late generation, Vigna unguiculata


Des études en champs ont été conduites pour évaluer la ségrégation des familles en F3 (jeune génération) et F6 (génération avancée) quant au poids des graines, la longueur de la gousse et le nombre de jours à la floraison chez des sous-espèces de niébé croisées. Une large gamme de lignées ségrégantes était introduite dans ces croisements et les familles utilisées étaient sensiblement différentes pour les trois caractères agronomiques étudiés. Les distributions continues observées pour ces caractères étudiés au sein des deux générations confirment la nature quantitative du patrimoine héréditaire pour ces caractères. Les estimations générales d'héritabilité se rangeaient entre 47,8 et 91,1%. Les estimations de transferts génétiques allaient des plus basses aux plus élevées et elles étaient constantes au sein des deux générations pour tous les caractères étudiés. Les corrélations entre générations s'est rangées entre 0.35 et 0.49 et montrées aussi les fortes associations entre les caractères mesurés pour les deux générations. Un déclin non significatif était observé entre la moyenne de F3 et son correspondant de F6. Ceci suggère l'existence d'une bonne mesure additive et peut-être de l' additive x additive composante de la variance (fixable pendant le croisement ultérieur) même si certains traits de dominance et de double épistasies (non fixables) pourraient être opérationnels. Les résultats de cette étude montre qu'une sélection parmi les jeunes générations de races supérieures est faisable.

Mots Clés: Caractéristiques agronomiques, jeune génération, génération avancée


Cowpea (Vigna unguiculata L.Walp) is a highly self-pollinated crop and the procedures in use for cultivar development have followed the conventional methods of individual plant selection in naturally occuring or hybridisation-induced genetic variability, following the pedigree method of breeding (Allard, 1960). An important assumption underlying early generation selection generally adopted for self-pollinated species is that selection for a character in the early generation (F2 or F3) would be as effective as when practised in the later generations assuming high heritability (Bartley and Weber, 1952; Allard, 1960). The use of early generation testing to identify superior crosses and eliminate large amounts of material from a cultivar development programme may increase breeding efficiency (Knauft and Wynne, 1995). An early and accurate appraisal of segregates has been of vital interest to most breeders of self-pollinated species including soybean, barley, peanut, oats and wheat; and early generation testing studied extensively in some allogamous species, has suggested the possibility of a similar application in autogamous species (Mahmud and Kramer, 1951; Knauft and Wynne, 1995). More precisely, Mahmud and Kramer (1951) have indicated that the two closely related problems are first, a selection of those crosses which are most likely to give the highest proportion of superior segregates, and second, an evaluation of the potentialities of the segregates from those crosses. Selection of crosses has been attempted on the basis of their performance in tests of bulk hybrid populations. It has been shown that replicated tests of segregating populations in the F2 or F3 generation would provide the average yield performance of the different crosses (Fehr, 1987).

Seed weight, pod length (Nakawuka and Adipala, 1998) and days to flowering are important agronomic traits in cowpea. Insight into the genetic segregation of these traits could be of benefit in the choice of methods to enhance selection efficiency. Several published quantitative studies on the genetics of these traits have been reported (Fery, 1985; Fery and Singh, 1997). Pedigree selection is commonly used in cowpea breeding programmes. However, comparisons among the various methods for selection in early generations and advancement to near-homozygous lines have not been made in cowpea.
The present study was conducted to determine the relationship of F3 lines to individual F6 lines derived by single-seed descent from the F3 lines of an inter-sub-specific cowpea cross. An attempt was made to avoid a possible interaction of seasons and generations by testing the F3 and F6 lines in the same year.


A cross was made between an improved cowpea line, IT84s-2246-4 (Vigna unguiculata ssp. unguiculata) and a wild relative, TVNu110-3A (Vigna unguiculata ssp. dekindtiana var. Pubescens), and advanced to the F2 generation. A plant from each F3 family was advanced to F5 by single-seed descent (SSD) method to produce the F6 seeds. The two parental genotypes and their F3 and F6 families were established in a field trial that was conducted to evaluate the performance of the F3 and F6 generations at the International Institute of Tropical Agriculture (IITA), Ibadan, Nigeria (7o30/ N; 3o54/ E) in late August, 1996. The IITA, is located at about 243 m above sea level with a mean annual temperature of 22.03°C (min.) and 30.33°C (max), mean seasonal rainfall of 1600 mm, and relative humidity of 88% during the late growing season (late August to late November) (IITA, unpublished).

The F3 and F6 plants and parents were grown in three replicates in a randomised complete block (RCB) design. Single-row plots were used for each entry per replicate. Each entry consisted of 16 plants maintained at 1 plant/hill (after thinning) at a spacing of 0.5 m within and 1.0 m between the rows. A total of 100 F3 families derived from the 100 F2 plants, and 94 F6 families derived by SSD were used for this study. Necessary agronomic practices were carried out to ensure optimum crop growth. Data on three important agronomic traits were recorded for each F3, F6 and the parental lines, as follows:

(i) Days to 50 percent flowering: the number of days from planting to when 50 percent of the plants had flowered.

(ii) Pod length: the length of fully mature pods, indicated by the change in pod colour, measured to the nearest centimetre. The average length of 40 pods was used as the score for each family in each replication.

(iii) 100-seed weight: weight in grams of 100 seeds after oven drying to a uniform moisture level of ca.13.5%. The average weight determined from four replicate samples was used as the score for each family in each replication.

Lines, generations, and replications were considered random effects in the combined analysis of variance over generations. A variance components (VARCOMP) procedure of the statistical analysis system (SAS) was used to calculate the maximum likelihood (ML) estimates of broad-sense heritability (hb) for the agronomic traits. Environmental variance was estimated as the mean square for the replicate x family interaction (Anderson et al., 1991). Genetic advance (GA) was calculated according to the formula used by Johnson et al. (1955):

GA (%) = ____ X 100

where k is the selection differential. It was given a value of 2.06, which is its expectation at 5% selection intensity from a normally distributed population, σ2g is the genotypic variance, x is the family mean, and σp is the phenotypic standard deviation.
The phenotypic coefficient of variation (PCV) and genotypic coefficient of variation (GCV) were calculated according to the formula used by Empig et al. (1970) as follows:

v σ2 p v σ2 g
PCV = ____ x 100 and GCV = ____ x 100 x x

where σ2p and σ2g are phenotypic and genotypic variances, respectively, and x is the family mean.


High variability was observed in the parents and segregating populations with respect to days to flowering, pod length and 100-seed weight. The mean squares from the combined analysis of variance over generations for these agronomic traits are presented in Table 1. Families were highly significantly (P<0.01) different for all three characters studied. The generations x families interaction, which is primarily a measure of environmental variation, was highly significant (P<0.01). No significant difference was observed between the F3 and F6 generations among all the three characters.

The frequency distributions for each of the three characters studied in this cross showed continuous variation, suggesting polygenic inheritance for all the traits. Histograms showing seed weight, pod length and days to flowering in the F3 and F6 generations are shown in Figures 1, 2 and 3, respectively. The population means and the mid-parental values are also shown for each trait in both generations. Days to flowering ranged from 46 to 63 and 44 to 65 days; pod length ranged from 6.3 to 10.0 and 5.5 to 12.0 cm; and seed weight ranged from 1.9 to 7.9 and 2.4 to 7.4 g/100 seeds in the F3 and F6 generations, respectively. The F6 progenies did not differ in their mean values from the F3 lines in all the traits studied. None of the segregants had seeds as heavy as the heavier seed-producing parent (IT84S-2246-4) in all generations (Fig. 1). Individual segregants exhibiting early flowering equal to that of the early flowering parent (IT84S-2246-4) were observed in all generations; but segregants with late flowering equal to that of the late flowering parent (TVNu110-3A) were not observed (Fig. 2). The population means were closer to the early flowering parent (IT84S-2246-4) in all generations (Fig. 2), suggesting partial dominance of early flowering. No individual segregant had pods as long as the longer podded-parent (IT84S-2246-4). However, individual segregants that were shorter than the short podded-parent (TVNu110-3A) were observed (Fig. 3). In both generations, the population means were closer to the length of the short podded-parent (TVNu110-3A) (Fig. 3), suggesting partial dominance of short pod. The values in all generations were nearer those of the parent with lower seed weight, suggesting partial dominance of small seed weight.

The means, standard errors and broad sense heritability estimates for the characters studied in this cross are shown in Table 2. Broad sense heritability estimates were moderate to high depending on the trait with values ranging from 47.8%, 74.9% and 91.1% for days to flowering, pod length and seed weight, respectively.

Estimates of phenotypic coefficient of variation (PCV), genotypic coefficient of variation (GCV) and genetic advance (GA) expressed in percent for traits in the F3 and F6 generations are shown in Table 3. The PCV and GCV values ranged from 10.6 to 46.1% and 5.4 to 25.5%, respectively, for characters in the F3 generation and were close to the estimated values in the F6 generation. The estimated GA (%) was consistently high for 100-seed weight in both generations; and low for days to flowering. A fairly high GA was observed for pod length. Associations between performance of F3 and F6 progenies were determined by intergeneration correlation coefficients and are presented in Table 4. All coefficients were significant at 1% probability level.


The cross used in this study provided a large range of genotypes and each differed in extent of variation for all attributes. Families were highly significantly different for all the traits studied. No significant differences are observed between the F3 and F6 generations in all the three characters studied. Inter-generation correlation coefficients for all the traits studied in the F3 and F6 generations were highly significant. Thus, there appears to be no important genetic reason why early generations should not provide good estimates of average yield potentialities of segregates from segregating populations when genetic shifts and interactions with environmental factors are controlled.

Analysis of the traits studied in the F3 and F6 generations of this cross showed continuous distributions, indicative of quantitative patterns of inheritance. The mean values of both generations for days to flowering were closer to the value of the early flowering parent, suggesting partial dominance of the alleles for early flowering. Adu-Dapaah et al. (1988) have observed a tendency for dominance of early flowering in cowpea. The mean values for pod length and seed weight in the four generations were closer to the short-podded and small-seeded parents, respectively, suggesting partial dominance of the alleles for these traits. This observation on seed-weight is consistent with previous reports in several plant species including cowpea (Drabo et al., 1984; Fatokun et al., 1992; Maughan et al., 1996). Nienhuis et al. (1987) reported a similarly skewed distribution for seed-weight that favoured the alleles of the wild parent in an interspecific cross of tomato.

Effective selection for desired traits when conditioned by quantitative inheritance is usually difficult in segregating populations and real progress from selection may be difficult to achieve. Hence, it may become desirable to determine genotypic variance in a segregating population in order to know the magnitude of the heritable fraction. A knowledge of such genetic parameters as heritability, phenotypic coefficient of variability (PCV), genotypic coefficient of variability (GCV), genetic advance (GA) and association among quantitative traits should help improve the efficiency of selection.

The extent of variability in these agronomically important traits has been well demonstrated by phenotypic and genotypic coefficients of variation. The highest GCV was observed for 100-seed weight in both generations while the lowest was observed for days to 50% flowering. This can be explained by the fact that segregating progeny were scored for days to 50% flowering on a progeny-row basis. Consequently, late flowering segregates obscured the early segregates in the row with a resultant bias toward lateness. This fact may also account for the low heritability estimates obtained for this trait.

Permanent gain from selection depends on the degree of relationship between genotype and phenotype. The correspondence between genotype and phenotype of a trait is expressed by heritability of the trait, which largely reflects the extent to which genetic segregation is expected in later generations of a cross for the trait in question (Mahmud and Kramer, 1951). Lush, 1948 (cited in Weber and Moorthy, 1952) defined heritability in two ways, broad sense heritability and narrow sense heritability. In the broad sense, heritability refers to the ratio of heritable variance to total variance. In the narrow sense, heritability is defined as the ratio of additive genetic variance to total variance. Broad sense heritability estimates in this study ranged from moderate to high. The heritability values obtained in this cross are within the values reported from several recently published studies in cowpea (Fery, 1985; Fery & Singh, 1997; Nakawuka & Adipala, 1998) and these estimates are encouraging for selection of families for the traits studied.

Heritability indicates the effectiveness with which selection of genotypes can be based on phenotypic performance. If heritability were 100% (i.e., σ2g=σ2p) then phenotypic performance would be a perfect indication of genotypic value; but even in such a hypothetical situation, the heritability value in itself provides no indication of the amount of genetic progress that would result from selecting the best individuals. According to Johnson et al. (1955) heritability estimates along with estimates of genetic advance (GA) were more useful than heritability alone in predicting the resultant effect for the selection of the best individuals from segregating populations. Very high heritability and GA (%) observed for 100-seed weight indicates that additive genetic variance is important for this trait and improvement can be achieved for this trait by phenotypic selection. The estimated GA (%) was consistently high in the F3 and F6 generations, indicating that the superiority of lines selected on the basis of F3 data may be retained in the later generations. Moderate heritability coupled with low GA (%) observed for days to 50% flowering indicates that little progress can be made in the improvement of this trait by phenotypic selection. Though high heritability was observed for pod length, estimates of genetic advance were only fairly high. No significant drop was observed between F3 mean and the corresponding F6 mean which may suggest the existence of a good measure of additive and possibly of additive x additive components of variance (which alone are fixable through subsequent inbreeding) although some amount of dominance and duplicate epistasis (which are non-fixable) may also be operative. The results of this study indicate that selection in early generations in cowpeas for superior types is feasible.


We thank Dr.C.A.Fatokun, and Prof. M.E. Aken'Ova and Dr. I .Fawole for their immense contributions.


  • Adu-Dapaah, H., Singh, B. B., Chheda, H. R. and Fatokun, C. A. 1988. Heterosis and inbreeding depression in cowpea. Tropical Grain Legume Bulletin 35:23-27.
  • Allard, R. W. 1960. Principles of Plant Breeding. New York: JohnWiley and Sons. 485 pp.
  • Anderson, W.F., Holbrook, C. C. and Wynne, J. C. 1991. Heritability and early generation selection for resistance to early and late leafspot in peanut. Crop Science 31: 588-593.
  • Bartley, B. G. and Weber, C. R. 1952. Heritable and non-heritable relationships and variability of agronomic characters in succesive generations of soybean crosses. Agronomy Journal 44:487-493.
  • Drabo, I., Redden, R., Smithson, J. B. and Aggarwal, V. D. 1984. Inheritance of seed size in cowpea [Vigna unguiculata (L.) Walp.]. Euphytica 33:929-934.
  • Empig, L., Lantican, T. and Escuro, P. B. 1970. Heritability estimates of quantitative characters in Phaseolus aureus
  • Robx. Crop Science 10:240 - 242.
  • Fatokun, C.A., Menancio-Hautea, D. I., Danesh, D. andYoung, N. D. 1992. Evidence for orthologous seed weight genes in cowpea and mungbean based on RFLP mapping. Genetics 132:841-846.
  • Fehr, W. R. 1987. Principles of cultivar development: Theory and Technique. Macmillan Publishing Company, NY. 536pp.
  • Fery, R.L. 1985. The genetics of cowpeas: A review of the world literature. In: Cowpea Research, Production and Utilization. Singh, S.R. and Rachie, K.O. (Eds.), pp. 25-62. John Wiley and Sons Ltd.
  • Fery, R.L. and Singh, B. B.1997. Cowpea genetics: a review of the recent literature. In: Advances in Cowpea Research, IITA/JIRCAS Co-publication. Singh, B. B., Mohan, D. R. Raj, Dashiel, K. E. and Jackai, L. E. N. (Eds.), pp. 13-29. IITA, Ibadan.
  • Johnson,H.W., Robinson, H. F. and Comstock, R. E. 1955. Estimates of genetic and environmental variability in soybeans. Agronomy Journal 47:314-318.
  • Knauft, D. A. and Wynne, J. C. 1995. Peanut breeding and genetics. Advances in Agronomy 55:393-445.
  • Mahmud, I. and Kramer, H. H. 1951. Segregation for yield, height, and maturity following a soybean cross. Agronomy Journal 43:605-609.
  • Maughan, P. J., Saghai Maroof, M. A. and Buss, G. R. 1996. Molecular-marker analysis of seed weight: genomic locations, gene action, and evidence of orthologous evolution among three legume species. Theoretical and Applied Genetics 93:574-579.
  • Nakawuka, C.K. and Adipala, E. 1999. A path coefficient analysis of some yield component interactions in cowpea. African Crop Science Journal 7:327-331.
  • Nienhuis, J., Helentjaris, T., Slocum, M., Ruggero, B. and Schaefer, A. 1987. Restriction fragment length polymorphism analysis of loci associated with insect resistance in tomato. Crop Science 27:797-803.
  • Weber, C.R. and Moorthy, B.R. 1952. Heritable and non-heritable relationships and variability of oil content and agronomic characters in the F2 generation of soybean crosses. Agronomy Journal 44:202-209.
© Copyright 2001, African Crop Science Society

The following images related to this document are available:

Photo images

[cs01065t4.jpg] [cs01065t1.jpg] [cs01065f1.jpg] [cs01065f2.jpg] [cs01065f3.jpg] [cs01065t2.jpg] [cs01065t3.jpg]
Home Faq Resources Email Bioline
© Bioline International, 1989 - 2022, Site last up-dated on 11-May-2022.
Site created and maintained by the Reference Center on Environmental Information, CRIA, Brazil
System hosted by the Internet Data Center of Rede Nacional de Ensino e Pesquisa, RNP, Brazil