M. S. Mohammed Z. Russom S. D. Abdul

Transcription

1 DOI /s Inheritance of hairiness and pod shattering, heritability and correlation studies in crosses between cultivated cowpea (Vigna unguiculata (L.) Walp.) and its wild (var. pubescens) relative M. S. Mohammed Z. Russom S. D. Abdul Received: 9 June 2008 / Accepted: 12 October 2009 Ó Springer Science+Business Media B.V Abstract Crosses were made between five cultivated cowpea varieties and a wild relative var. pubescens (as pollen parent) to study the inheritance of hairiness, pod shattering, as well as heritability and correlations among yield and yield related traits. F 1 plants exhibited dominance for both hairiness and pod shattering traits. Segregation in F 2 populations approximated 3:1 and 13:3 for hairy and non-hairy stem and 13:3 for hairy and non-hairy pod. The ratio of 13:3 and 15:1 ratios were observed among the F 2 populations for shattering and non-shattering pods. Broadsense heritability estimates showed that days to pod maturity averaged 77.93%, 100 M. S. Mohammed (&) Z. Russom S. D. Abdul Crop Production Programme, School of Agriculture and Agricultural Technology, Abubakar Tafawa Balewa University, Bauchi, Nigeria sagirkura@yahoo.com Z. Russom rusoom17@yahoo.co.uk S. D. Abdul sdanabdul2003@yahoo.com S. D. Abdul Biological Science Programme, School of Science and Science Education, Abubakar Tafawa Balewa University, Bauchi, Nigeria Present Address: M. S. Mohammed Department of Plant Science, Institute for Agricultural Research, Samaru, Ahmadu Bello University, Zaria, Zaria, Nigeria seed weight 68.45%, seeds pod % and number of branches plant -1 had 62.54% (all high). Days to first flowering and pod length were moderate (57.31 and 54.29%, respectively). Number of pods plant -1 had low heritability estimates with an average of 39.0%. Correlations among number of pods plant -1, seeds pod -1, seeds plant -1 and seed yield plant -1 ( ) were significantly positive. Correlations between seed pod -1 and seed plant -1 with number of pods plant -1 were and 0.996, respectively. High broadsense heritability averaged 77.93% for days to pod maturity, 68.45% for 100 seed weight, 69.76% for seeds pod -1 and 62.54% for number of branches plant -1.Daysto flowering and pod length plant -1 had moderate estimates of and 54.29%, respectively, while pods plant -1 was low (39.90%). Potential utilization of the hairiness trait and the implication of pod shattering in cowpea improvement were discussed. Positive correlations among these yield components, as well as high to moderate broadsense heritability estimates of the yield related traits can be utilized for direct and indirect selection to improve grain yield in cowpea. Keywords Broadsense heritability Cowpea Correlation F 1 Hairiness Pod shattering Introduction Cowpea is a major crop world wide, producing a source of economic livelihood and nutritional

2 well-being for millions of farmers and urban consumers in the developed and developing world (Timko 2006). Cowpea currently ranked 23rd among important crop species which is grown on over 14.5 million ha world wide (IITA 2004), and with an average annual grain production of 12 million tones of which more than two-third (about 7.6 million tones) is produced in sub-saharan Africa (AATF 2006). Cowpea has several advantages. The grain contains between 25 and 32% protein (Bressani 1985) making it extremely valuable where many people cannot afford animal protein. It is cultivated in the semi-arid regions of lowland tropics and sub-tropics where the soil is poor in fertility and rainfall is scanty (Mortimore et al. 1997). Cowpea is shade tolerant and therefore can be intercropped with a number of cereals, root crops, cotton, banana and plantain (Alghali 1991). In an intercropping, cowpea provides ground cover, and this in turn conserves moisture, suppresses weeds, and it may also provide some protection against soil erosion (Quin 1997). Also, some cowpea genotypes induce suicidal germination of Striga hermonthica which may parasitize the associating cereal (Blade et al. 1997). Symbiosis by root nodule bacterial (Bradyrhizobium species) increases the fertility level of the soil (Blade 1993). The tender leaves and green pods are used as vegetables. Cowpea haulms are used as fodder for cattle. However, cowpea is primarily a grain food legume. The dry seeds are boiled and eaten with rice or alone. They are also processed and eaten as moi-moi (steamed paste) and bean cake (fried paste) as well. Despite all the aforementioned advantages, the major drawback of cowpea is its low yield, mainly due to among other factors, severe attacks from extensive pest complexes (Rachie 1985). Most elusive among cowpea insect pests are cowpea legume pod borer (Maruca vitrata) and a complex of pod sucking bugs both of which can cause high seed yield losses (Jackai 1995; Fatokun et al. 1997). Intercropping as a component of integrated pest control (IPC) does not necessarily reduce the pest load in any given situation, but largely depends on the associating crop(s) and the pest(s) in question (Jackai and Adalla 1997). It is however, noteworthy that none of the companion crops significantly reduces cowpea damage by among other pests, M. vitrata and the podsucking bug complex, all of which constitute the major pests of cowpea (Emeasor and Ezueh 1997). Upon all these, chemical insecticide approach seems to be the most effective mean of controlling these pests but with consequent residual effects on the user and environmental pollution as well. Therefore, it is clear that availability of cowpea varieties with resistance to these pests would be attractive to cowpea producers so that the crop could be grown with less dependence on expensive, often adulterated chemicals that are not particularly environmentally friendly (Fatokun 2000). Wild relatives of many crop species including cowpea are often found to be sources of important genes for disease and insect resistances, increased yield, improved product quality, earliness and wide crop adaptation (Leppik 1970; Stalker 1980). Accessions of wild Vigna species including var. pubescens (TVNu 110-3A) were screened and found to confer some degree of insect resistance on cowpea (Fatokun et al. 1997; Fatokun and Singh 2001). The reason for the insect resistance could be associated to presence of hairs (pubescens) on the plants. Thus, it will be of interest to transfer the hairiness trait from the wild Vigna species to the cultivated cowpea varieties. Unfortunately, var. pubescens was found to possess undesirable gene(s) for pod shattering: the opening of mature pods along both the dorsal and ventral sutures in the cowpea. Presently, there has been scanty information on the mode of inheritance of hairiness and pod shattering traits in cowpea, because the traits are not important among cultivated cowpea, but found most frequently in wild species including var. pubescens. Cultivated cowpea varieties and var. pubescens (TVNu 110-3A accession) belong to section Catiang, but different subspecies. However, several reports have shown that members of section Catiang are cross compatible; hence exchange of gene(s) should not be difficult to accomplish (Fatokun et al. 1997). In crop breeding for yield improvement, breeders commonly consider several characters and their relationships as well as their predictive confidence (repeatability) as base information for the evaluation of cultivars. They often use selection indices, which form a method for the quantitative integration of such information. The selection procedure simultaneously weighs the genetic heritability and correlations of characters of interest for the improvement of crops. Such selection goes with the relative economic value that provides phenotypic index for evaluation of cultivars.

3 Quite often, characters are correlated; and selection for one character may lead to negative or positive response in the other character (Ajibade and Morakinyo 2000). However, selection for certain characters could depend on the magnitude of their heritability estimates. Estimation of heritability and its interpretation has presented many challenges to plant scientists (Nyquist 1991). Correlation studies are of great value in determining the association of qualitative characters with yield for selecting the characters which played important role that influence yield in cowpea. In the light of these, this work was carried out statement of objectives rewritten to capture reviewers comment #1: to confirm the genetic crossability between five cultivated cowpea varieties and their wild relative var. pubescens (TVNu 110-3A) as well as the reproductive potential of the F 1 hybrids, the inheritance of stem and pod hairiness, and pod shattering in the crosses. Broadsense heritability estimates and the relationships between some agronomic and yield, and yield related traits in the crosses as well as their implication in cowpea improvement were reported. Materials and methods Screen house and field experiments were conducted at the Abubakar Tafawa Balewa University, Bauchi. Bauchi is in the northern Guinea savannah of Nigeria, located N and E, and 609 m above sea level with a rainfall of mm per annum. Materials used in this study were four improved cultivated cowpea varieties and one non-cultivated wild relative, var. pubescens (TVNu 110-3A) obtained from the International Institute of Tropical Agriculture (IITA) Kano sub-station, Kano, Nigeria. The improved cultivated cowpea varieties were: IT89KD-288, IT93K-452-1, IT97K , and IT81D-994 while Dandunga a local cowpea variety, also cultivated, from Bauchi was included as check. The screen house experiment involved hand crosses between the five cultivated cowpea varieties and TVNu 110-3A. This exercise was carried out from 10 to 29 September, 2003, as described by Myers (1991) with some modifications. TVNu 110-3A was used as pollen parent and crossed to each of the five cultivated cowpea varieties. F 1 seeds were harvested at maturity, and were later advanced to F 1 plants to obtain F 2 seeds. Seeds each of the six parents, the five F 1 and F 2 genotypes were sown in rows in the field during 2004 rainy season in a randomized complete block design with three replications. Each replicate had 16 plots. The plots measured 4.5, 4.5 and 58.5 m 2 for each of the six parents, five F 1 s genotypes and five F 2 s genotypes, respectively. A plant was maintained to each stand and spaced 75 cm 9 1 m within and between rows, respectively. Twelve plants were thus maintained to each plot of the parents and the F 1 genotypes, while each of the F 2 populations had 108 plants per plot. Cultural practices observed include land clearance, ploughing and harrowing using tractor. No ridges were made, so planting was done on flat. The plots were weeded twice starting at four weeks after sowing followed by a second weeding three weeks later using hand hoes. Casual hand picking of broad leaved weeds was also carried out as weeds emerge. Number of plants bearing hairs on either stems or pods or both, as well as those plants without hairs on both stems and pods were recorded. Presence of hairs on both the stems and the pods was assessed manually by hand feeling on all the plants across the entire experiment. For pod shattering, pods were scored as either shattering or non-shattering across the entire experiment. The results were subjected to Chi-square (v 2 ), to test for goodness of fit. For the determination of broadsense heritability and correlations among yield and yield related traits, the following data were recorded as follows: Days to first flowering Determined by counting the number of days from sowing to the day the first flower appeared for each plant in each plot. Days to 95% pod maturity Determined by counting the number of days from sowing to the day when about 95% of the pods on each plant mature. Number of pods plant -1 The actual number of mature pods harvested on each plant. Pod length Determined by measuring (with ruler) the lengths of 10 randomly selected pods for each plant, from the tip to the point of attachment on the peduncle and the mean obtained. Number of seeds pod -1 This was determined by taking average number of seeds from the 10 randomly selected pods used to measure pod length for each genotype. Hundred (100) seed weight This was determined by counting and taking weights of 100 seeds

4 after counting number of seeds pod -1 for each genotype. Number of seeds plant -1 This was initially calculated by multiplying the number of pods plant -1 with the number of seeds pod -1. Number of branches This was determined by counting total number of branches for each plant from which average number of branches were estimated for each of the 16 genotypes. All the recorded data were subjected to analysis of variance. These as well as the derived data were subjected to correlations coefficient using the following formula: P n i¼1 r ¼ x iy i 1 P n n i¼1 x P n i i¼1 y i qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi P n i¼1 x2 i 1 P n n i¼1 x qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 P n i i¼1 y2 i 1 P n n i¼1 y 2 i where r correlation coefficient; P summation; x and y independent variables to be compared (if x and y are as stated here, the formula may not be true cf. end of the document); P xy sum of the xy or summation xy; P x 2 sum of squared variable x or summation x squared variable; and P y 2 sum of squared variable y or summation y squared/sum of squared variable x squared variable. Days to first flowering, days to 95% pod maturity, number of branches plant -1, pod length, number of pods plant -1, number of seeds pod -1 and 100 seed weight were further subjected to estimation of broadsense heritability (H b ) as described by Kelly and Bliss (1975) using the following formula: H b ¼ S2 F 2 S 2 P 1 þ S 2 P 2 þ S 2 F 1 =3 S F 2 where H b heritability in the broadsense; S 2 P 1 and S 2 P 2 variances of parents 1 and 2, respectively; and S 2 F 1 and S 2 F 2 variances of F 1 and F 2 generations, respectively. Results and discussion Crossability between cultivated cowpea and the wild relative The five cultivated cowpea varieties crossed well with the hairy wild relative var. pubescens (TVNu 110-3A) via conventional hand pollination. A total of 800 flowers were emasculated and pollinated out of which 432 formed mature pods, representing 54% of pod set. An average of 54% of pod set arising from hand crossing using the wild type as pollen source compared favourably with 50% arising from natural crossing among the parent genotypes (Tables 1, 2). Also, about 64 73% of the flowers in the F 1 crosses formed mature seeds and each of the five crosses had very high percentage of mature pods than either of the two parents. This clearly shows that there was high parent heterosis for pod set in these crosses and with these parents, and hence devoid of any incompatibility problems. Seeds of the five F 1 crosses were viable thus suggesting good reproductive potential (Table 2). Stem hairiness, pod hairiness and pod shattering Presence and absence of hairs on the stem and pod among the parent lines, F 1 and F 2 crosses are presented in Tables 3 and 4. All the cultivated parents were pure for the absence of hairs on the stem and the pod, while the wild parent had hairs on both the stem and the pod. However, all the five F 1 crosses had hairs on their stem and their pod, indicating dominance of the gene(s) for stem and pod hairiness, respectively. The result further showed that, the hairiness trait on both positions (stem and pod) segregated independently among the F 2 populations, but with some peculiarities. It is pertinent to state that this is one of the very few reports of genetic Table 1 Number of flowers pollinated, number of pod set and the percentage of pod set in crosses involving cultivated and wild (V. unguiculata var. pubescens) cowpea varieties Cross No. of flowers pollinated No. of pod set Percentage pod set (%) Dandunga 9 wild type IT89KD wild type IT93K wild type IT97K wild type IT81D wild type Total (average)

5 Table 2 Mean number of flowers per plant and percentage (%) of flowers that produced mature pods per plant among parents and F 1 s in crosses involving cultivated and wild (V. unguiculata var. pubescens) cowpea varieties Means followed by the same letter(s) within the same column are not significantly different at 5% level of probability (DMRT) Genotype Mean number of flowers per plant No. of pod per plant % Mature pods Parent Dandunga 23d 16e 69.6 IT89KD d 16e 55.2 IT93K cd 13e 39.4 IT97K d 13e 46.4 IT81D cd 20e 62.5 Wild (var. pubescens) 122a 23cd 29.5 Mean F 1 cross Dandunga 9 wild type 66b 48abc 72.7 IT89KD wild type 60bc 44abcd 73.3 IT93K wild type 86b 55a 64.0 IT97K wild type 64b 40bcd 62.5 IT81D wild type 76b 50ab 65.8 Mean SE ± (0.05) inheritance of hairiness in cowpea. F 2 population of Dandunga 9 wild cross segregated in the ratio of 13:3 for the presence and absence of hairiness on both the stem and the pod (Tables 3, 4). This suggests the interaction of dominant and recessive genes for the inheritance of the trait. The 3:1 ratio observed in the remaining crosses (IT89KD wild, IT93K wild, IT97K wild and IT81D wild) for stem hairiness suggests a monogenic inheritance with dominance for hairiness (Table 3). However, these same crosses had 13:3 ratios each for the inheritance of the hairiness on the pod (Table 4), which could be due to epistatic interaction of dominant and recessive genes. The segregation pattern observed in these crosses for pod hairiness could be as a result of interaction between two independently segregating gene loci that show epistatic gene action. In the work of Aliyu (2007) involving crosses to study the inheritance of pubescens density and length on the leaves, stems and pods using a different wild species, similar monogenic and digenic segregation ratios among the F 2 populations were obtained. Hence, it was observed in this work that even though both the homozygotes and the heterozygotes are not readily distinguishable among the F 2 populations of all the crosses, inheritance of the trait can be selected for in early generations, in the case of 3:1 ratio. Similarly, selection where digenic interaction is involved with epistatic action as reported here can be possible in the early generations. Results for pod shattering (Table 5) showed that all the cultivated parents had intact pods, while pods of the wild parent and the five F 1 genotypes shattered suggesting dominance of pod shattering over nonshattering. Similar report was made by Aliboh et al. (1997) in their study involving crosses between cultivated cowpea and two wild and weedy spp. They further showed that using the two wild spp. in the crosses they made, all F 2 populations had a ratio of 3:1, indicating monogenic dominance of pod shattering over non-shattering in their own report. In contrast, this study showed that F 2 populations of Dandunga 9 wild, IT93K wild and IT81D wild relative (var. pubescens) segregated into 13:3 ratios for the inheritance of pod shattering in cowpea (Table 5). This segregation pattern suggested the association of both dominant and recessive genes. IT89KD wild and IT97K wild relative segregated into 15:1 ratio for shattering over non-shattering pods, suggesting the involvement of duplicate dominant genes for the inheritance of the trait. Differences between this study and reports of Aliboh et al. (1997) may be due to differences in the genotypes of the wild ssp., or even the cultivated cowpeas used.

6 Table 3 Segregation pattern for stem hairiness of the parents, F 1 s and F 2 sin crosses involving cultivated and wild (V. unguiculata dekindtiana var. pubescens) cowpea varieties Genotype Number of plants Expected ratio v 2 Probability Hairy Non-hairy Parent Dandunga 0 35 IT89KD IT93K IT97K IT981D Wild (var. pubescens) 31 0 F 1 cross Dandunga 9 wild type 35 0 IT89KD wild type 34 0 IT93K wild type 34 0 IT97K wild type 34 0 IT981D wild type 36 0 F 2 cross Dandunga 9 wild type : IT89KD wild type : IT93K wild type : IT97K wild type : IT981D wild type : Table 4 Segregation pattern for pod hairiness of the parents, F 1 s and F 2 sin crosses involving cultivated and wild (V. unguiculata dekindtiana var. pubescens) cowpea varieties Genotype Number of plants Expected ratio v 2 Probability Hairy Non-hairy Parent Dandunga 0 35 IT89KD IT93K IT97K IT981D Wild (var. pubescens) 31 0 F 1 cross Dandunga 9 wild type 35 0 IT89KD wild type 34 0 IT93K wild type 34 0 IT97K wild type 34 0 IT981D wild type 36 0 F 2 cross Dandunga 9 wild type : IT89KD wild type : IT93K wild type : IT97K wild type : IT981D wild type :

7 Table 5 Segregation pattern for pod shattering of the parents, F 1 s and F 2 sin crosses involving cultivated and wild (V. unguiculata dekindtiana var. pubescens) cowpea varieties Genotype Number of plants Expected ratio v 2 Probability Shattering Non-shattering Parent Dandunga 0 35 IT89KD IT93K IT97K IT981D Wild (var. pubescens) 31 0 F 1 cross Dandunga 9 wild type 35 IT89KD wild type 34 IT93K wild type 34 IT97K wild type 34 IT981D wild type 36 F 2 cross Dandunga 9 wild type : IT89KD wild type : IT93K wild type : IT97K wild type : IT981D wild type : Broadsense heritability Days to first flowering and 95% pod maturity Broadsense heritability estimates for days to first flowering and days to 95% pod maturity in this study were low to high ( %) with an average of 57.31% and moderate to high ( %) with an average of 77.93%, respectively (Table 6) among the crosses. Fery and Singh (1997) showed high broadsense heritability estimates of 75 and 79%, for days to 50% flowering and 95% pod maturity, respectively. Ramachandran et al. (1980) and Roquib and Patnaik (1990) also made similar reports. When traits or characters are highly heritable in the broadsense (i.e. more than 70%), it means that such traits or characters could be selected for in the early generation assuming there is significant additive gene action. Number of branches Estimates of broadsense heritability were found to be moderate to high ( %) for number of branches plant -1 with an average of 62.51%. It was observed in this study that genotypes with more number of branches relatively performed better in terms of yield components such as number of seeds plant -1 and total seed yield. Although these genotypes matured later, such results have advantage while selecting genotypes with high H b estimates as obtained in the crosses IT93K wild type (71.43%) and IT97K wild type (83.54%). Reports were made on broadsense heritability estimates for number of branches in cowpea with variations among studies. Fery (1985), and Fery and Singh (1997) reported broadsense heritability estimates of 45 and 57% (moderate), respectively for number of branches plant -1 in cowpea. Apte et al. (1987) reported broadsense heritability estimates of 34.9% (low) for number of branches plant -1 in cowpea. Although results obtained from this study were in similar range with earlier reports, few of the estimates lower than what was observed may be due to the use of wide crosses in this study. Pod length and number of seeds pod -1 Heritability estimates for pod length was low to highly heritable ( %), with an average of 54.29%. The H b estimate for IT89K wild type

8 Table 6 Broadsense heritability (H b ) estimates in crosses involving cultivated and wild cowpea varieties Genotype (cross) and H b values (%) Average a P 1 P 6 P 2 P 6 P 3 P 6 P 4 P 6 P 5 P 6 H b values (%) Days to maturity Days to pod maturity Number of branches plant Pod length No. of pods plant No. of seeds pod Hundred (100) seed weight a P 1 = Dandunga, P 2 = IT89KD-288, P 3 = IT93K-452-1, P 4 = IT97K , P 5 = IT981D-994, P 6 = V. unguiculata dekindtiana var. pubescens (wild) (-8.76%) means there is no heritability. Negative estimates of H b are best explained in terms of environmental influence, sampling error during data collection or, that one or both parents are heterozygous in those loci for the trait in question. Therefore, our study suggests that the negative H b could be due to the heterozygocity of the cultivated parents involved in the crosses since it was not observed in other crosses involving other cultivated parents. However, considering these implications, selection of genotypes for cowpea yield improvement should not be based on pod length. Particular reference needs to be made of IT97K wild type, which had H b estimate of 79.19% (high). However, possible reasons for the negative H b estimate in Dandunga 9 wild type ( %) for number of seeds pod -1 could be due to sampling error, or heterozygocity of one or both parents at loci controlling this trait. Studies on pod length by Fery (1985) indicated moderate to high heritability estimates up to 75.2%. However, this author reported moderate estimate (52.8%) for number of seeds pod -1, while 83.0% (high) heritability estimate for number of seeds pod -1 was reported (Roquib and Patnaik 1990). Number of pods plant -1 Broadsense heritability estimate was low to moderate for number of pods plant -1 across all the crosses in this experiment. The range was % with an average of 39.90%. On a general note, if not for IT93K wild type which had moderate (46.08%), H b estimates for the trait, in all the crosses, were low. Similar observation of low broadsense heritability estimates of 20 and 19.25% was reported (Ajibade and Marakinyo 2000). Negative H b estimates in IT89KD wild type and IT81D wild type i.e and %, respectively, could be due to sampling error, heterozygocity of one or both parents for the loci controlling this trait, or environmental influence. The implication of low H b estimates for number pods suggests that, singular selection for high yield in cowpea cannot be based on number of pods plant -1 especially if the trait is not associated with additive gene effect. Broadsense heritability estimates of 53.1% (Fery 1985); 7.0% (Bliss et al. 1973); 21.4% (Trehan et al. 1970) and 85.9% (Damarany 1994) were reported for number of pod plant -1. Hundred (100) seed weight Broadsense heritability estimates for hundred seed weight were mostly found to be high in this study ( %) with an average of 68.43%. In similar observations, Damarany (1994) reported 83.3%, while 96% was obtained by Ajibade and Morakinyo (2000). IT89KD wild type had a moderate estimate of 42.86%. The correlation and heritability values and magnitude reported in crosses between cultivated and improved cowpea varieties and their wild relative were in most cases in the same range as reported by earlier workers. Therefore, other yield performances as well as heritability and correlations values obtained in this study should be considered and articulated for yield improvement programmes in cowpea. The implication of low H b estimates for

9 number pods plant -1 suggests that, singular selection for high yield in cowpea cannot be based on number of pods plant -1. However, the interrelationship between number of pods plant -1 and other yield components which were found to be highly heritable can be utilized for yield improvement in cowpea. Such correlations were observed in seeds pod -1, seeds plant -1 and total seed yield. Although broadsense heritability estimates for 100-seed weight was high in this study, results of correlations indicated inverse relationship of this trait with some of the yield components. This suggests that high H b estimates for 100-seed weight on one hand and correlations of yield components on the other hand should be a priority, while selecting genotypes for yield improvement in cowpea. The estimates could hence be useful information especially if the trait is associated with additive genes effect. Lopes et al. (2003) reported that five genes control the expression of the inheritance of seed size in cowpea which are additive. Correlations Number of days to first flowering was positive and highly correlated with days to 95% pod maturity (0.96). Also number of days to first flowering was significantly and positively correlated with number of pods plant -1 (0.16), number of seeds plant -1 (0.23) and seed yield plant -1 (0.15) (Table 7). Furthermore, number of days to 95% pod maturity was significantly and positively correlated with number of pods plant -1 (0.21), number of seeds pod -1 (0.31) and number of seeds plant -1 (0.21). The interrelationships of these characters could lead to indirect selection of one or more of these characters that are relevant to yield improvement. High and positive correlations between seed yield and pods plant -1, days to 50% flowering and plant dry weight were reported (Aggarwal 1986). Number of seeds pod -1 showed high and significant positive correlations with number of seeds plant -1 and yield plant -1 (0.94 and 0.57, respectively), while seed yield plant -1 was significantly and positively correlated with number of seeds plant -1 (0.53) and 100- seed weight (0.10). Similar report was made by Singh et al. (1982). This means that indirect selection of these characters to improve yield could be effective in crosses between cultivated and improved cowpea and their wild relative. However, pod length was inversely correlated with days to first flowering (-0.15), days to 95% pod maturity (-0.24), number of pods plant -1 (-0.63), number of seeds pod -1 (-0.54) and number of seeds plant -1 (-0.59). Again, 100-seed weight was negatively correlated with all characters, except pod length and yield plant -1. The implication of inverse relationship between or among yield components and agronomic traits such as days to first flowering and days to 95% pod maturity means that, indirect selection of days to first flowering and/or 95% pod maturity and 100-seed weight could lead to reduction in yield. Number of pods plant -1 was significantly and positively correlated with number of seeds pod -1 (0.91), number of seeds plant -1 (0.99), and seed yield plant -1 (0.49). This is in agreement with the report of Ajibade and Morakinyo (2000) as mentioned earlier. This means the selection of these yield components could be utilized to improve yield in cowpea. Noteworthy were, the near perfect positive correlations between pods plant -1 and seeds pod -1, pods Table 7 Correlation coefficients among agronomic, yield and yield-related characters in crosses involving cultivated and wild cowpea Days to first flowering Days to maturity Pod length Pods plant -1 Seeds pod -1 Seeds plant seed weight Days to maturity 0.963** Pod length * * Pods plant * 0.210* ** Seeds pod * 0.305* ** 0.909** Seeds plant * 0.210* ** 0.996** 0.935** 100 seed weight * * 0.886** ** ** ** Yield plant * 0.485* 0.568** 0.529** 0.097* * P \ 0.05, ** P \ 0.01

10 plant -1 and seeds plant -1 and, seeds pod -1 and seeds plant -1. This is expected because a character such as seeds plant -1 was derived from the other characters mentioned above. However, this trend could be utilized for the direct and indirect selection of these characters to improve yield in cowpea. Conclusion Wild relatives of cowpea such as var. pubescens are sources of useful genes for incorporation into the cultivated cowpea (Mohammed 2007, Unpublished M.Sc. Thesis). However, the inherent pod shattering trait of these wild relatives is an undesirable character in the cultivated cowpea. The fact that the mode of inheritance of pod shattering was dominant in the F 1 and controlled by digenic genes in the F 2 segregating populations, suggests the likelihood of difficulty in eliminating the trait. Although there were variations in the estimates of heritability in the broadsense among the genotypes with respect to the characters investigated in this study, characters with high heritability estimates can be utilized in cowpea breeding programme for yield improvement. This is because characters with high heritability can easily be transferred from parents to offsprings and hence are inherited within early generation of selection. The near perfect positive correlation between number of pods plant -1 and seeds pod -1, number of pods plant -1 and seeds plant -1, and seeds pod -1 and seeds plant -1 could be utilized for the direct and indirect selection to improve yield in cowpea. It is therefore suggested that the potential of var. pubescens is effectively utilized while the undesirable traits such as pod shattering are eliminated. Acknowledgements We wish to acknowledge the management of the Kano State Agricultural and Rural Development Authority, Kano Nigeria, and the Institute for Agricultural Research Samaru, Ahmadu Bello University Zaria, Nigeria, for partial funding of the first author s M.Sc. program; Senior author s M.Sc. programme and Prof. M. P. Timko, Department of Biology, University of Virginia, USA for his academic suggestions. References African Agricultural Technology Foundation (AATF) (2006) Accessing proprietary technology for improving cowpea productivity. Annual Report, 38 pp Aggarwal VD (1986) Cowpea breeding. In: IITA, Grain Legume Improvement Program, pp Annual Report 1985 IITA/SAFGRAD Ajibade SR, Morakinyo JA (2000) Heritability and correlation studies in cowpea (Vigna unguiculata [L] Walp.). Niger J Sci 15:29 33 Alghali AM (1991) Studies on cowpea farming practices in Nigeria with emphasis on insects pest control. Trop Pest Manag 37:71 74 Aliboh VO, Kehinde OB, Fowale I (1997) Inheritance of leaf mark, pod dehiscence and dry pod colour in crosses between wild and cultivated cowpeas. Afr Crop Sci J 5(2): Aliyu B (2007) Heritability and gene effects for incorporating pubescence into cowpea (Vigna unguiculata (L.) Walp.) from V. rhomboidea Burtt. Davy. Euphytica 155: Apte UB, Chavan SA, Jadhav BB (1987) Genetic variability and heritability in cowpea. Ind J Agric Sci 57: Blade SF (1993) Summaries of review papers. Crop Improvement Division P IITA Ibadan, Nigeria Blade SF, Shetty SVR, Terao T, Singh BB (1997) Recent developments in cowpea cropping systems research. In: Singh BB, Mohan Raj DR, Dashiell KE, Jackai LEN (eds) Advances in cowpea research. IITA/JIRCAS IITA, Ibadan, Nigeria, pp Bliss FA, Barker LN, Franckowiak JD, Hall TC (1973) Genetic and environmental variation of seed yield, yield components and seed protein quantity and quality of cowpea. Crop Sci 13: Bressani R (1985) Nutritive value of cowpea. In: Singh SR, Rachie KO (eds) Cowpea research, production and utilization. Wiley, New York, pp Damarany AM (1994) Estimates of genotypic and phenotypic correlation, heritability and potency of gene set in cowpea (Vigna unguiculata [L.] Walp.). Assuit J Agric Sci 25:1 8 Emeasor KC, Ezueh MI (1997) The influence of companion crops in the control of insect pests of cowpea in intercropping systems. Trop Agric 74: Fatokun CA (2000) Breeding cowpea for resistance to insect pests: attempted crosses between cowpea and Vigna vexillata. In: Fatokun CA, Tarawali SA, Singh BB, Kormawa PM, Tamo M (eds) Challenges and opportunities for enhancing sustainable cowpea production. IITA, Ibadan, Nigeria, pp Fatokun CA, Singh BB (2001) Improving cowpea-cereals systems in the dry savannas. Annual Report 2001 IITA, Ibadan, Nigeria, 79 pp Fatokun CA, Perrino P, Ng NQ (1997) Wide crossing in African Vigna species. In: Singh BB, Mohan Raj DR, Dashiell KE, Jackai LEN (eds) Advances in cowpea research. IITA/JIRCAS IITA, Ibadan, Nigeria, pp Fery RL (1985) Improved cowpea cultivars for the horticultural industry in the USA. In: Singh SR, Rachie KO (eds) Cowpea research production and utilization. Wiley, New York, pp Fery RL, Singh BB (1997) Cowpea genetics: a review of the recent literature. In: Singh BB, Mohan Raj DR, Dashiell KE, Jackai LEN (eds) Advances in cowpea research. IITA/JIRCAS IITA, Ibadan, Nigeria, pp International Institute of Tropical Agriculture (IITA) (2004) Improving and intensifying cereal legume systems in

11 moist and dry savannas of West and Central Africa. Annual Report, IITA, Ibadan, Nigeria, 170 pp Jackai LEN (1995) The legume pod borer, Maruca vitrata and its principal host plant Vigna unguiculata (L.) Walp. Use of insecticide sprays as an aid in the identification of useful levels of resistance. Crop Prot 14: Jackai LEN, Adalla CB (1997) Pest management practices in cowpea a review. In: Singh BB, Mohan Raj DR, Dashiell KE, Jackai LEN (eds) Advances in cowpea research. IITA/JIRCAS IITA, Ibadan, Nigeria, pp Kelly JD, Bliss FA (1975) Heritability estimates of percentage seed protein and available methionine and correlations with yield in dry beans. Crop Sci 15: Leppik EE (1970) Gene centers of plants as source of disease resistance. Annu Rev Phytopathol 8: Lopes FCC, Gomes RLF, Filho FRF (2003) Genetic control of cowpea seed sizes. Sci Agricol 60(2):1 8 Mortimore MJ, Singh BB, Harris F, Blade SF (1997) Cowpea in traditional cropping systems. In: Singh BB, Mohan Raj DR, Dashiell KE, Jackai LEN (eds) Advances in cowpea research. IITA/JIRCAS, IITA, Ibadan, Nigeria, pp Myers GO (1991) Hand crossing of cowpea. IITA Research Guide No. 42. IITA, Ibadan, Nigeria, 18 pp Nyquist WE (1991) Estimation of heritability and prediction of selection response in plant populations. Crit Rev Plant Sci 10: Quin FM (1997) Introduction. In: Singh BB, Mohan Raj DR, Dashiell KE, Jackai LEN (eds) Advances in cowpea research. IITA/JIRCAS IITA, Ibadan, Nigeria, pp ix xv Rachie KO (1985) Introduction. In: Singh SR, Rachie KO (eds) Cowpea research, production and utilization. Wiley, New York, pp xxi xxviii Ramachandran C, Peter KV, Gopalakrishnam PK (1980) Variability in selected varieties of cowpea (Vigna unguiculata [L.] Walp.). Agric Res J Kerala 18(1):94 97 Roquib MA, Patnaik RK (1990) Genetic variability in forage yield and its components in cowpea Vigna unguiculata [L.] Walp. Environ Ecol 8(1): Singh R, Joshi BS, Singh S (1982) Correlation studies in cowpea (Vigna unguiculata [L.] Walp.). In: Tropical grain legume bulletin 26. International Institute of Tropical Agriculture, pp 3 4 Stalker HT (1980) Utilization of wild species for crop improvement. Adv Agron 33: Timko MP (2006) The cowpea genomics initiative, 2nd edn. Department of Biology, University of Virginia, USA Trehan KB, Bagrecha LR, Srivastava VK (1970) Genetic variability and correlation in cowpea (Vigna unguiculata [L.] Savi.) under rain fed conditions. Ind J Hered 2(1): 39 43