Evaluation of some Musa accessions in field collection at Ebonyi State University , Abakaliki , Nigeria

The study was conducted to characterize and evaluate the agronomic potentials of eight Musa accessions in Ebonyi State University Musa field genebank. Qualitative descriptors were used to characterize the genotypes, while eight growth and yield traits were evaluated. Results indicated that the accessions were similar in thirteen characteristics, but differed in nine traits including pseudostem/foliage pigmentation, pseudostem/foliage waxiness, pseudostem blotching, petiole margin clasping, peduncle hairiness, persistence of neutral flowers, persistence of male bract, pulp colour of unripe fruit, and pulp colour of ripe fruit. Analysis of variance results indicated significant differences among the accessions for all the agronomic traits except fruit circumference. In terms of bunch weight, SH 3436 yielded the heaviest bunch (12.45 kg), followed by FHIA 25 (10.03 kg) and FHIA 17 (8.43 kg), respectively. These three genotypes were all banana hybrid varieties and may have adapted well to the climatic conditions of Abakaliki in southeast agro-ecological zone.


INTRODUCTION
Bananas and plantains (Musa species) are important staple foods in the humid tropics of the world and are sources of revenue for smallholder farmers (Vuylsteke et al., 1993a).Bananas and plantains are cultivated in compound or backyard gardens that are rich in organic matter and nutrients from household refuse.These gardens are continuously cultivated with plants growing vigorously, giving an annual yield of about 30 to 50 tonnes per hectare (Nweke et al., 1988).The fruits are high in dry-matter content, with long shelf life and good cooking qualities that make them attractive to consumers.The fruits are highly nutritious, containing large amounts of carbohydrates and minerals such as phosphorus, calcium, and potassium as well as vitamins A and C (Sharrock and Frison, 1998).Fruit production occurs throughout the year in consecutive cycles, which ensures year round food security and income for farmers (Njuguna et al., 2008).
Thus, it is important to conserve and maintain the genetic diversity of Musa spp., notably for use in breeding *Corresponding author.E-mail: happinessoselebe@yahoo.com.
Author(s) agree that this article remain permanently open access under the terms of the Creative Commons Attribution License 4.0 International License new lines, generating planting materials for farmers, sustaining future and extended agricultural production and ensuring food security.The Musa field genebank in Ebonyi State University (EBSU) was established for this purpose.
There is wide morphological variation among banana (AAA genomes) and plantain (AAB genomes) cultivars for bunch type, plant size, fruit orientation, fruit shape, size of the pseudostem and fruit colour (Simmonds and Shepherd, 1955;Tezenas du Montcel et al., 1983;Swennen, 1990).Four types of plantains have been distinguished based on inflorescence morphology, namely, French plantain, French Horn plantain, Horn and False Horn plantain (De Langhe, 1964;Tezenas du Montcel and Devos, 1978).De Langhe (1964) was the first to classify the plantain cultivars according to the size of the pseudostem into giant, medium and small types.Plant size reflects the number of leaves produced prior to flowering, potentially indicating differences in photosynthetic capacity of plants (Stover and Simmonds, 1987).To better understand the uniqueness of the accessions and varieties in EBSU Field germplasm, they needed to be characterized and evaluated.
Germplasm characterization is an important operation for a gene bank.The value of the germplasm collection depends on the availability of information relative to the accessions.Morphological and agronomic traits as well as reaction to biotic and abiotic stresses that are known to be in the individual accessions determine their relative value characterization and evaluation; they also assist in maintaining the genetic integrity of plant materials after an extended period of continuous propagation, and help germplasm users to recognize clones with desirable traits for later use in crop improvement programs.
The objectives of the study were to (a) characterize accessions in the Musa genebank of Ebonyi State University and (b) to evaluate these accessions for growth and yield characteristics.

MATERIALS AND METHODS
The study was conducted at the Musa field genebank, Department of Crop Production and Landscape Management, Ebonyi State University (EBSU), Abakaliki.Eight Musa accessions made up of four plantain and four banana accessions were used for the study.The plantain accessions included PITA 14, PITA 17, PITA 25 (developed by International Institute for Tropical Agriculture (IITA), Ibadan) (Vuylsteke et al., 1993b) and a local landrace called "Agbagba".The banana accessions included FHIA 25, FHIA 17 and SH 3436 (developed by Fundación Hondureña de Investigaçion Agricóla (FHIA) in Honduras) and a landrace denoted as "Lagos banana".Except 'Lagos banana' that was collected from a compound at Ikoyi, Lagos State, Nigeria, the other materials were sourced from the High Rainfall Station of IITA at Onne, Rivers State.
The field was established with minimum disturbance to the soil by adopting the zero-tillage land preparation method (Pars et al., 1990).Planting was done in 2006 following an unbalanced randomized complete block design whereby the collected Oselebe et al. 947 accessions were assigned into non-replicated 2 row plots, each with 5 plant stands.Planting holes were prepared at a minimum size of about 30 × 30 × 30 cm 3 .Suckers of acquired accessions of similar sizes were prepared for planting following procedures outlined by Swennen (1990).Prepared suckers were placed in the hole and their corm covered first with topsoil and subsequently with sub-soil.They were planted using a spacing of 3 × 2 m 2 to give a population density of 1,667 plants per hectare.
The field was maintained across years and evaluated in 2008.Mulching was done using grasses and dead leaves as well as fresh leaves and pseudostems of harvested genotypes.Weeds were controlled chemically at intervals using glyphosate and sometimes by manual weeding especially at the base of the mat.N: P: K 12:12:17 +2% MgO fertilizer was applied to each stand at the rate of 230 kg/ha.Pruning was done to eliminate young suckers except the tallest one to take over as the next ratoon after the harvest of the mother plant and to maintain plant density.Deleafing was also done regularly to ensure a neat plantation as well as good quality fruits.

Characterization of Musa germplasm
The Musa genotypes studied were characterized using qualitative descriptors as shown in Table 1.

Agronomic evaluation of field established genotypes
Data were collected for growth and yield attributes of the Musa genotypes as follows: (1) Days to fruit filling (DFF), that is, the number of days elapsed from the emergence of the inflorescence to time of harvest of the bunch when at least a finger from the first two proximal hands had shown sign of ripening; (2) Plant height at flowering (PHT, cm) measured from soil level to the point where the two highest petioles meet each other; (3) Height of the tallest sucker at harvest (HTS, cm) measured as in PHT; (4) Bunch weight (BWT, kg), that is, weight of a mature inflorescence of the plant carrying the fingers, grouped in hands.This was determined by weighing the bunch at harvest using a weighing balance; (5) Total number of fruits per bunch; (6) Number of hands per bunch; (7) Number of fruits per hand; (8) Fruit length (FTL) and fruit circumference (FRC): Fruit characteristics were determined from the middle fruit of the second hand and were expressed in cm.

Data analysis
Statistical analysis of data collected for agronomic parameters was done using the general linear model (GLM) procedure in Statistical Analysis System (SAS), SAS Institute (2000), Version 9; using the linear additive model for analysis of variance in RCBD experiment, thus: where Xij = individual observation of i th treatment in the j th block; µ = overall mean of the population; τi = the effect of the i th treatment; Bj = Block effect; and Eij = experimental Error (residual) per block.Significant parameters were subjected to mean separation test, using the Least Significant Difference (LSD) procedure at P < 0.05,
while simple linear correlation was conducted based on mean values of the accessions to draw inferences on trait associations.

Characterization of Musa accessions
Results obtained using the qualitative descriptors indicated that the accessions were similar in thirteen characteristics, but different in nine traits classified under vegetative, inflorescence and fruit traits.

Texture of male bracts
Persistence o f male bract PITA 14, PITA 17, FHIA 17, and SH 3436.Only "Lagos banana" had persistent male bracts (that is, the male bracts dried-off on the male axis), while others had deciduous male bracts.

Agronomic evaluation of Musa accessions
Results of analysis of variance (Tables 5 and 6) indicated very highly significant differences (P < 0.001) among the Musa accessions for days to fruit filling, plant height at flowering, number of hands per bunch and total number of fruits per bunch.It also showed highly significant difference (P < 0.01) among the accessions for bunch weight and average fruit length.The accessions were significantly different (P < 0.05) for height of tallest sucker at harvest and average number of fruits per hand.Conversely, there was no significant difference among the accessions for fruit  circumference.

Mean performance of eight Musa accessions grown at Ebonyi State University for some phenological traits
The results showed that PITA 25 had the highest number of days to fruit filling (111.83days), which differed significantly from all the other accessions studied except "Agbagba" the plantain landrace (104 days).All the banana accessions ("Lagos banana", FHIA 25, FHIA 17, and SH 3436) had lower number of days to fruit filling compared to the plantains, and values obtained were not significantly different from each other (Table 7).The lowest number of days to fruit filling was recorded for "Lagos banana" (approximately 92 days).

Yield components of Musa accessions
SH 3436 produced the biggest bunches (12.45 kg) which differed significantly from bunch weight for PITA 14, PITA 17, PITA 25, Agbagba, and "Lagos banana".This was followed by FHIA 25 (10.03kg), which significantly differed from PITA 17, "Agbagba" and "Lagos banana", only.The bunch weights of the rest of the accessions were statistically the same.Results also indicated that FHIA 25 and FHIA 17 produced the highest number of hands in a bunch, having mean number of hands approximating 8 and 7, respectively.They differed significantly from PITA 14, PITA 17, PITA 25, Agbagba, and "Lagos banana" on this trait.PITA 17 produced the least number of hands per bunch (approximately 4 hands) that was statistically the same with the number of hands recorded for PITA 14, PITA 25, "Agbagba" and "Lagos banana".In terms of average number of fruits per hand, FHIA 25 had the highest number of fruits (13.0), followed by FHIA 17 and SH 3436 (approximately 12.5 and 12,respectively).This significantly differed from number of fruits per hand for PITA 14, "Agbagba" and "Lagos banana"."Lagos banana" produced the least number of fruits per hand (9.0) that was statistically the same with PITA 14, PITA 17 and "Agbagba".Among the accessions, PITA 25 produced the longest fruits (21.2 cm) that differed significantly from the fruits of FHIA 17, FHIA 25, PITA 17, and "Lagos banana".This was followed closely by SH 3436 (19.98 cm)."Lagos banana" produced the shortest fruits (14.6 cm) that was statistically the same with fruit length for PITA 17, FHIA 17 and FHIA 25, but differed from fruits of SH 3436, PITA 14, PITA 25 and "Agbagba".Conversely, FHIA 25 produced the highest number of fruits per bunch (approximately 100).This differed significantly from the number of fruits produced by PITA 14, PITA 17, PITA 25, "Agbagba" and "Lagos banana", but not from number of fruits produced by FHIA 17 and SH 3436 (approximately 87 and 81, respectively).PITA 14 produced the least number of fruits per bunch (approximately 42) in this experiment.

Phenotypic correlation among traits
Results indicated significant positive correlation (P <0.05) between days to fruit filling, plant height at flowering, height of tallest sucker at harvest and length of fruits (Table 8).A significant and positive correlation (P< 0.001) was observed between plant height at flowering and height of tallest sucker at harvest, bunch weight, average number of fruits per hand, average fruit length, and total number of fruits per bunch.Significant positive correlation (P < 0.01) existed between plant height at flowering, number of hands per bunch and fruit circumference.Again, strong positive correlation existed between bunch weight and its components, including number of hands per bunch, average number of fruits per hand, average fruit length, total number of fruits per bunch and fruit circumference.

DISCUSSION
To be of value, genotypes in a gene bank must be characterized.Characterization descriptors must exhibit polymorphism, either between or within taxa (Ortiz, 1997).Also, they should be highly heritable, easy to visually score and consistently expressed in all environments.The genotypes studied in this experiment exhibited differences in some vegetative traits, including level of pseudostem/foliage pigmentation, pseudostem/foliage waxiness, pseudostem blotching, and petiole margin clasping.The high level of red pigmentation observed in Agbagba in this study was as a result of the presence of anthocyanin (Ortiz, 1997).On the other hand, the absence of anthocyanin in "Lagos banana" may have resulted in the absence of pigmentation in this accession.
Anthocyanin presence was scored during the rainy season as dry weather tends to reduce pigmentation.In this experiment, some of the accessions possessed extensive pseudostem blotching, some possessed moderate blotching, some exhibited extensive pseudostem/foliage waxiness, while others were slightly waxy.Ortiz (1997) earlier reported that tetraploid bananas had drooping leaves and extensive blotching and waxiness of the pseudostem, corroborating our observations on two banana accessions (SH 3436 and "Lagos banana") in Ebonyi State University Musa field genebank.
All the genotypes were the same in leaf orientation, pseudostem colour, plant juice colour, sucker orientation and petiole canal.Leaves were drooping, pseudostems were green in colour, the plants contained white sap, petiole canal were enrolled and suckers emerged from the soil in an erect position parallel to the mother stem.For inflorescence traits, only two banana accessions ("Lagos banana" and FHIA 17) had coarsely hairy peduncles.Similar result was reported by Ortiz (1997) on Highland bananas.Accessions of PITA 14, PITA 17, PITA 25, "Agbagba", FHIA 25, and SH 3436 possessed finely hairy peduncle and were similar to those of ABB and AAA triploids as reported by Simmonds (1966).Neutral flowers abscised from the inflorescence stalk of PITA 14, PITA 17, FHIA 17, and SH 3436, whereas those of PITA 25, "Agbagba" and "Lagos banana" retained neutral flowers.Only "Lagos banana" clones retained dried male bracts on the male axis, while others had deciduous male bracts.The genotypes did not differ in bunch compactness, presence or absence of male axis, external colour of male bract, and texture of male bract as shown.
As stated by Medlicott et al. (1992), the fruit peel colour is often the major post-harvest criterion used by researchers, growers and consumers to determine whether the fruit is ripe or unripe.In some countries (e.g.Ghana, Nigeria, Honduras, etc.), consumers have developed distinct associations between colour and the overall quality of specific products.Cooking banana and plantain should be green or yellow and anything that falls short (e.g.red plantain) would be difficult to sell.Hence, colour is critical as the first visual assessment of the quality of cooking banana or plantain.In some West African countries, if the pulp colour of plantain or cooking banana is white, consumers feel that the fruit is immature and if the pulp colour is orange/yellow, it indicates that the fruit is mature.The disappearance or loss of peel green colour and the corresponding increase in yellowing of the peel during ripening are the obvious manifestations in banana, cooking banana and plantain.The loss of green colour is due to degradation of the chlorophyll structure.External changes in peel colour during ripening often reflect changes in pulp colour also (Wainwright andHughes, 1989, 1990) As observed in this study, peel colours of all the clones of the accessions were green when unripe and yellow when ripe.The pulp colours of plantain clones were orange/yellow when unripe, while those of banana clones were white.When ripe, the pulp of plantain clones turned yellow, while those of banana clones were creamcoloured.These observations are similar to the aforementioned peel and pulp colour characteristics for Musa species and as such, the market values of the fruits were not reduced by peel and pulp colour.
Plant height at flowering and height of tallest sucker at harvest were significantly different among the accessions studied.FHIA 17 was the tallest at flowering, while PITA 25 produced the tallest suckers at harvest.Héber and Ricardo (2001) reported similar results and indicated differences in these traits among two Musa cultivars (Common dwarf and Huamoa) studied in Puerto Rico.They also observed significant differences in bunch weight, number of fruits per bunch, and number of days to fruit filling among genotypes studied.Similar results were obtained in this study where, bunch weight, days to fruit filling, number of hands per bunch, number of fruits per hand and average fruit length were significantly different among the accessions indicating variation for the traits.Generally, the banana accessions required shorter number of days to fruit filling and their fruits attained the mature green stage in significantly shorter periods, compared to the plantain accessions.
The matrix of correlations between inflorescence and vegetative traits of the Musa accessions studied indicated that bunch weight was significantly associated with the number of hands and fruits per bunch.Consequently, the higher the number of hands and fruits per hand, the heavier the bunch becomes (Swennen et al., 1995).Bunch weight was also significantly associated with other traits such as average number of fruits per hand, total number of fruits per bunch, average fruit length, fruit circumference, plant height and height of the tallest sucker at harvest.
A positive significant correlation between days to fruit filling and plant height at flowering indicated that taller plants required more days for their fruits to be completely filled than shorter ones.An increase in plant height at flowering resulted in an increase in the number of days to fruit filling in the eight Musa genotypes evaluated in this study.In the same vein, a positive significant correlation was observed between number of hands per bunch and average number of fruits per hand, total number of fruits per bunch, average fruit length and fruit circumference signified that an increase in the number of hands per bunch resulted in a corresponding increase in average number of fruits per hands, total number of fruits per bunch, average fruit length, and fruit circumference.

Conclusion
This study established that the eight Musa accessions in Ebonyi State University field genebank were similar in some qualitative traits and different in others.Agronomic evaluation indicated accessions with high yield potentials, including, which may be utilized as parents in crosses where feasible.The high yield of some of the accessions SH 3436 and FHIA 25, may also indicate their adaptability within the agro ecological zone and may be recommended for cultivation as an increase in the farmer"s income is assured.

Table 1 .
Qualitative descriptors used in characterising accessions in Ebonyi State University Musa field germplasm.

Table 2 .
Characterization of eight Musa accessions based on vegetative traits.

Table 3 .
Characterization of eight Musa accessions based on inflorescence traits

Table 4 .
Characterization of eight Musa accessions based on fruit traits.

Table 5 .
Result of analysis of variance (mean square values) on the performance of eight Musa accessions grown at Ebonyi State University for some phonological traits.

Table 6 .
Result of analysis of variance on the performance of eight Musa accessions grown at Ebonyi State University for some yield components *, **, and ***Significant at 0.05, 0.01, and 0.001 probability levels, respectively.ns: Non-significant; DF: degrees of freedom.

Table 7 .
Mean performance and LSD values of eight Musa genotypes grown at Ebonyi State University for some yield components.

Table 8 .
Linear correlation coefficients among traits in eight Musa genotypes evaluated.