AGRONOMIC EVALUATION OF BANANA PLANTS IN THREE PRODUCTION CYCLES IN SOUTHWESTERN STATE OF BAHIA

Rev. Bras. Frutic., v. 39, n. 1: (e-990) DOI 10.1590/0100-29452017 990 Jaboticabal SP ABSTRACT This assay was conducted in the experimental area of the Federal Institute of Bahia, Campus Guanambi, BA, and aimed to evaluate agronomic traits of Prata, Cavendish, Gros Michel and Maçã banana cultivars three production cycles. The 72 treatments, 24 cultivars and three production cycles were arranged in a split plot scheme in time, in a completely randomized design with five replications and four plants per plot. Plots were arranged in 24 cultivars, Prata-Anã, Maravilha, FHIA-18, FHIA-18 BRS, BRS Platina, JV42-135, Pacovan, Japira, PV79-34, Pacovan-Ken, Preciosa, Guarantida, Maçã, Caipira, BRS Tropical, BRS Princesa, YB42-03, YB42-07, YB42-47, Grande-Naine, Calypso, Buccaneiro, FHIA-23 and FHIA-17; and subplots consisted of three production cycles. Data obtained were submitted to analysis of variance. The average of the cultivars were grouped by Scott-Knott criterion (p<0.05) and production cycles compared by Tukey test (p<0.05). ‘JV42-235’, ‘Japira’ and ‘Pacovan-Ken’ cultivars had larger size and ‘Grande Naine’ had smaller size. ‘Prata-Anã’ cultivar had higher number of leaves at harvest, with leaf area index similar to the others. ‘BRS Platina’ cultivar is earlier at flowering and harvest. ‘Maravilha’, ‘BRS Platina’, ‘FHIA-23’, ‘BRS Tropical and BRS Princesa’ cultivars presented greater potential for use by farmers.


INTRODUCTION
Banana cultivation in southwestern Bahia and northern Minas Gerais is characterized by the cultivation of Prata-Anã cultivar (RODRIGUES et al., 2011).The main phytosanitary problems affecting banana plants in these regions are yellow sigatoka (MATOS; CORDEIRO, 2011) and the Panama disease, another fungal disease that limits 'Maçã' banana cultivation (SILVA et al., 2002;MOHANDAS et al., 2013).This limitation also occurs in some areas of 'Prata' banana cultivation, where this disease has become severe, especially in regions with more sandy soil (RODRIGUES et al., 2011).
Alternatives have been proposed by the Brazilian Program for Genetic Improvement of Banana, through the development of resistant cultivars that have been tested in several regions of the country.At Guanambi-BA, some studies have already been conducted to evaluate banana genotypes, as those performed by Donato et al. (2006Donato et al. ( , 2009Donato et al. ( , 2010)), Azevedo et al. (2010), Faria et al. (2010), Marques et al. (2011) and Sant 'Ana et al. (2012).
Studies of this nature are important because they provide information about new cultivars, stability, behavior and resistance to diseases in different environments, and consolidate and strengthen the recommendation of genotypes in the test phase.Therefore, it is essential to continue these studies, not only with the most cultivated subgroup, but also with others evaluated in several production cycles in order to obtain a greater number of cultivation options for farmers, ensuring the consistency and extrapolation of results.
Thus, the aim of this study was to evaluate the agronomic characteristics of Prata, Cavendish, Gros Michel and Maçã banana cultivars in three production cycles in southwestern Bahia.

MATERIAL AND METHODS
The experiment was carried out at the Federal Institute of Bahia, Campus Guanambi, in a typical dystrophic Red-Yellow Oxisol, A weak, medium texture, with average annual precipitation of 678 mm and mean temperature of 26 ºC.
At planting, micropropagated seedlings spaced 3.0 x 2.5 m were used.The implantation and the cultural treatments followed recommendations for the culture, according to Rodrigues et al. (2008).
The 72 treatments, 24 cultivars and 3 production cycles were arranged in a plot scheme subdivided in time in a completely randomized design, with five replicates and four useful plants per plot.Plots consisted of 24 cultivars, Prata type: Prata-Anã and Pacovan (AAB), Maravilha, FHIA-18, BRS FHIA-18, BRS Platina, Pacovan, Japira, Pacovan-Ken, Preciosa, Garantida and genotypes PV79-34 and JV42-135 (AAAB); maçã type: Maçã(AAB), Caipira (AAA), BRS Tropical, BRS Princesa and genotypes YB42-03, YB42-17 and YB42-47 (AAAB); Cavendish type: Grande-Naine (AAA);Gros Michel type: Calipso, Bucaneiro, FHIA-23and FHIA-17 (AAAA); and subplots, for the three production cycles: mother plant, daughter plant and granddaughter plant.Growth, yield, and cycle duration were evaluated.In the growth and yield measurements, standard phenotypic descriptors were defined in the Descriptors Manual (IPGRI, 1996;BRANDÃO et al., 2013): plant height, corresponding to pseudostem length, pseudostem perimeter at 30 cm from the soil, number of live leaves at flowering and harvest, length and width of the third leaf, total leaf area and the leaf area index.Total leaf area of the plant (TLA), expressed in m², was estimated according to Zuculoto et al. (2008).At harvest, bunch and hand mass, average hand mass, number of hands and number of fruits were determined.Cycle duration (in days) was also evaluated, determining the flowering period, period between flowering and harvest, and harvesting period.
Data obtained were submitted to analysis of variance.The averages of cultivars were grouped by the Scott-Knott criterion (p <0.05), and the production cycles were compared by the Tukey test (p <0.05).

RESULTS AND DISCUSSION
Flowering and harvesting characteristics were influenced by cultivars and cycle (Tables 1, 2, 3, 4), except for the flowering period (Table 5).
Plant height differed among cultivars with the formation of six groups in the first and third cycles, and of five groups in the second, by the Scott-Knott criterion (p <0.05) (Table 1).The highest cultivars in all cycles were Japira, Pacovan-Ken and genotype JV42-135, and the smallest ones, Grande Naine and Caipira, agreeing with Camolesi et al. (2012).Banana plants increased in height during cycles, which is consistent with literature (SOUZA et al., 2011a;Oliveira et al., 2013).Possibly, the third cycle is representative of the height of each cultivar, since as reported by Belalcázar Carvajal (1991)  The identification of groups of cultivars of different sizes in evaluation trials may support future recommendations of specific planting spacings, since the plant height interferes with the spacing, planting density and consequently light capture, directly related to photosynthesis, which determines productivity (AZEVEDO et al., 2010).In addition, higher height makes harvesting and cultural management difficult.It is an undesirable characteristic in regions with strong winds due to intense plant bending.
Pseudostem perimeter, evaluated at 30 cm from the soil, in the first two cycles, formed six groups (Table 1).'FHIA-23' presented the largest perimeter, 109.45 cm, in the first cycle.In the second cycle, this cultivar, together with 'FHIA-18', 'Prata-Anã' and 'FHIA-17' presented the highest averages.In the third cycle, with seven groupings, the highest averages were observed for 'FHIA-18', 'Maravilha', 'BRS Platina' and 'PV79-34'.In all cycles, the lowest performance was observed for Caipira cultivar, similar to that observed by Camolesi et al. (2012).As in the present study, Azevedo et al. (2010) observed that 'PV79-34', a Pacovan-type hybrid susceptible to Panama disease (RODRIGUES FILHO et al., 2014), expresses greater vigor and smaller size than 'Pacovan' and its hybrids 'Preciosa', 'Garantida', 'Japira' and 'Pacovan-Ken'.Possibly, its male genitor identified as 79 (AA) provides the greatest vigor and smallest size (AZEVEDO et al., 2010), but it does not provide resistance to Panama disease as M53 to its progeny (RODRIGUES FILHO et al., 2014).There were differences in the pseudostem perimeter among cycles, with size increase for all cultivars, except for FHIA-17, which did not present difference, and for 'FHIA-23', the only one that presented continuous size reductions over the cycles.
Increases in the pseudostem perimeter of banana plants from the first production cycle are reported in several studies, such as Damatto Júnior et al. ( 2011), with 'Prata-Anã' (AAB), Souza et al. (2011b), with 'Figo Cinza' (ABB) and Silva et al. (2013), with 'Thap Maeo' (AAB).According to Silva et al. (2011), this feature is of great importance for the culture, because the pseudostem perimeter is related to the ability of the bunch to support, susceptibility to bending, number of leaves and bunch development; therefore, it expresses the plant vigor.
Banana plants presented a number of leaves evaluated in flowering, in three groups, in the first and third cycles and four groups in the second cycle (Table 1).In the first cycle, averages ranged from 12 to 18 leaves for 'BRS FHIA-18' and 'Prata-Anã', respectively.In the second cycle, 'Prata-Anã' and 'Maçã' showed higher number of leaves, and the group with the lowest number included the majority of cultivars, with average around 12 leaves.In the third cycle, the majority of cultivars formed the group with the lowest number of leaves, while 'BRS Platina' and 'Pacovan' cultivars presented the highest number.The number of leaves in flowering did not differ among months for most cultivars.The differences observed indicate a reduction in this characteristic from the first to the other cycles.
The number of live leaves is an important characteristic to be analyzed, considering that the photosynthesis rate depends on the plant leaf area (OLIVEIRA et al., 2013).Souza et al. (2011a) did not find differences among cultivars of three genomic groups, Grande Naine, Caipira, Prata-Anã, FHIA-18 and BRS Tropical evaluated in subtropical climate.Different results were found in this work for most plants in the three cycles.From flowering, there is no emission of leaves, which decreases in number until harvest by the natural process of foliar senescence and defoliation to control sigatoka infestation (OLIVEIRA et al., 2013).
The number of leaves at harvest varied among cultivars in the evaluated cycles, forming three, six and four groups for the first, second and third cycles, respectively.'Prata-Anã' cultivar had the highest number of leaves in all cycles, along with 'Maçã', in the first and third cycles.The latter group also included 'Maravilha' and 'BRS Platina' cultivars.The lowest number of leaves in all cycles was observed in 'Garantida', 'Preciosa' and 'Caipira' cultivars.In the majority of variations observed among cycles, in 50% of the cultivars evaluated, there was a reduction in the number of leaves at harvest along the cycles.The number of leaves at harvest for Thap Maeo cultivar (AAB), with 'Maçã' type fruits during two production cycles in the Vale do São Francisco region (SILVA et al., 2013), was about 9 and 10, respectively, values similar to those obtained in this work.
The length of the third leaf showed a small difference among cultivars in the first and second cycles, forming three and two groupings by the Scott-Knott criterion (p <0.05) (Table 2).The highest leaf length was observed for 'FHIA-23' and 'FHIA-17' cultivars, and the smaller leaf lengths were observed for 'Prata' type fruits: 'BRS FHIA-18', 'BRS Platina', 'Prata-Anã'; 'Grande Naine' of the Cavendish group and 'Maçã' type fruits: 'Caipira' and 'BRS Princesa'.In the second cycle, a small variation occurred with formation of only two groupings, and banana plants with the shortest length of the third leaf were the same as those of the previous cycle.This feature reflects genotypic differences, since Gros Michel banana plants have larger leaves.Most cultivars increased the length of the third leaf from the first to the second cycle.
The width of the third leaf varied among cultivars, forming five groups in the first cycle (Table 2).The lowest leaf width was recorded for 'Maçã' cultivar, while the largest was observed for 'Calipso' and 'Bucaneiro', in both cycles, in the latter group 'FHIA-23', 'FHIA-17' and 'Grande Naine' were also grouped.This was expected due to the genetic proximity of these cultivars.In the majority of cultivars, the width of the third leaf among cycles did not differ, but there was a decrease from the first to the second cycle in 'Maravilha', 'Pacovan-Ken', 'Pacovan' and 'Caipira' cultivars.Marques et al. (2000) evaluated 'Prata-Anã' and 'BRS Platina' (PA42-44) cultivars in three production cycles with different irrigation systems in Guanambi-BA and found differences in the width of the third leaf only in the second cycle.'Prata-Anã' cultivar emitted wider leaves compared to 'BRS Platina' cultivar, with values ranging from 66.14 cm to 79.23 cm, a result similar to that obtained in this study.
The total leaf area varied similarly in both evaluated cycles, forming three groups (Table 2).In the first cycle, the majority of cultivars were included in the group with the highest leaf area, while the lowest results were obtained for 'BRS FHIA-18', 'Caipira', 'YB42-03' and 'BRS Princesa' cultivars.In the second cycle, the lower averages were found for 'Garantida' and 'Caipira' cultivars.There was a statistical difference between cycles, with increases in leaf area for 'BRS FHIA-18', 'Prata-Anã', 'BRS Tropical', 'BRS Princesa' and 'Maçã' cultivars, and decrease for 'Pacovan Ken' and 'FHIA-17' cultivars from the first to the second cycles.
The results of this experiment exceeded those of Costa et al. (2012), in which leaf area at flowering for Galil-18 (FHIA-18) cultivar ranged from 9.29 to 12.26 m 2 , being significantly influenced by irrigation depths in Cruz das Almas-BA.
Agronomic characteristics of 'Prata' type banana in three productive cycles evaluated by Marques et al. (2011) in Guanambi-BA, reached leaf area values around 9.75 and 11.70 m² for 'Prata-Anã' and 'BRS Platina' cultivars, respectively, lower than those observed in this study.
Cultivars presented the same leaf area index in both cycles with the formation of only one group by the Scott-Knott criterion (p <0.05) (Table 2).The leaf area index did not differ between cycles for most cultivars, and increases were observed for 'BRS FHIA-18', 'Prata-Anã', 'BRS Tropical', 'BRS Princesa' and 'Maçã' While for 'Pacovan-Ken' cultivars and decreased for 'FHIA-17' cultivar.
Superiority of leaf area indexes of 'Prata-Anã' cultivar when compared to its progeny, 'BRS Platina', in three cycles evaluated under different irrigation systems was also recorded by Marques et al. (2011).The results ranged from 1.30 to 2.36 m 2 m -2 , whose values are similar to those found in this study.The variation of the number of hands among cultivars was similar in the three cycles evaluated with the formation of three clusters (Table 3).Cultivars originating from the breeding program of Honduras (FHIA) presented the highest number of hands in the first two cycles.In the third cycle, 'Bucaneiro' cultivar stood out with the highest number of hands, 16.42.In all cycles, the smallest number of hands was registered in tall 'Prata' banana cultivar and in those with 'Maçã' type fruits, as well as in 'Calipso' cultivar.There was no difference between the cycles for most cultivars; however, among those that showed difference, BRS FHIA-18, 'BRS Platina', 'FHIA-18', 'Caipira' and 'YB42-03' cultivars presented lower values in the first cycle and higher in the third cycle, 'Bucaneiro', the only one that presented increase in the number of hands from the first to the second and from the second to the third cycle.
The superiority of 'Prata-Anã' cultivar in relation to the number of number of hands per bunch, when compared to 'BRS Platina' cultivar, observed by Marques et al. (2011) in three production cycles, was only observed in the second cycle of the present study.
The lowest number of fruits per hand was observed in tall 'Prata' banana plants in all cycles, and the highest in the 'Cavendish' and 'Gros Michel', and in 'Caipira' cultivars (Table 3).Among cases of variation, most of the cultivars showed an increase in the number of fruits per hand throughout cycles.Similar results were found by Silva and Rodrigues (2013) when evaluating bunch mass, number of hands and fruits per bunch of 'Prata-Anã' banana plants, regardless of application of P in the soil, in four production cycles.
'Maravilha' cultivar showed the highest average hand mass in the three cycles, while the lowest average mass was obtained for 'Maçã'-type banana plants, with exception of 'Tropical BRS' in the second cycle and 'YB42-17' in the third cycle (Table 3).There were differences among cycles with the reduction of the average hand mass from the first to the third cycle for most cultivars.The largest reduction was observed for 'FHIA-23' and 'FHIA-17' cultivars.
Hand mass varied among cultivars with formation of six, four and five groups in the first, second and third cycles, respectively (Table 3).Cultivars originating from FHIA reached higher values in all cycles, as banana plants with 'Maçã'type fruits presented lower hand mass in all cycles.There were no differences among cycles for banana plants with 'Pacovan' and 'Maçã' -type fruits.'FHIA-18', 'FHIA-18' and 'Bucaneiro' cultivars presented mass increment from the first to the second cycle and were similar in the third cycle, while 'FHIA-17' and 'FHIA-23' cultivars decreased from the second to the third cycle.
Fresh hand mass is directly related to mass and number of fruits; however, this positive correlation can be greater or lower, depending on the genotype and environment (LIMA NETO et al., 2003).Souza et al. (2011b) evaluated the first two cycles of the 'Figo Cinza' cultivar and verified an increase in the fresh mass values of the second hand, without finding an increase in the number of fruits.
Banana plants presented small variation among cultivars in relation to the number of days to flowering, with formation three groups in the first two cycles and greater variation with formation of five groups in the third cycle (Table 4).The highest precocity of flowering in all cycles was found for 'BRS  were the latest in all cycles, both in flowering and harvesting.Banana plants varied in number of days for flowering among cycles, obviously, this variation is explained by the fact that the number of days for flowering is cumulative from one cycle to another.
The flowering precocity is an important characteristic because it reduces the time of plant exposure to pathogens, being able to increase the number of leaves at the time of floral differentiation and favor a greater amount of female flowers during inflorescence (ROBINSON; GALÁN SAÚCO, 2010), resulting in bunches with greater number of hands.
Data measured by Azevedo et al. (2010) indicate that, during the period from planting to flowering, 'Garantida', 'PV42-53', 'Preciosa', 'Pacovan-Ken', 'Japira' cultivars and 'PV79-34' hybrids were later than 'Pacovan' in the first two cycles.Similar behavior was observed in the present experiment in the second and third production cycles, that is, 'Pacovan' was earlier than hybrids, except for 'PV79-34' in the second and third cycles.
The variation in the number of days to harvest among cultivars allowed four groups in the first and third cycles and three in the second (Table 4).The earliest cultivars in all cycles were BRS FHIA-18 and BRS Platina.
The reduction in the number of days required for the bunch emission brings expectations of return of the initial investment (FARIA et al., 2010).Melo et al. ( 2010) observed the anticipation of 103 days of harvest in the first cycle of 'Prata-Anã' cultivar when they used 1,016 kg ha -1 of N and 1,200 kg ha -1 of K 2 O together, and reported that the cycle of this cultivar is extended in N and K deficiency situations.
'Prata-Anã' cultivar cultivated under the edaphoclimatic conditions of Juazeiro-BA presented in the first production cycle an interval of 347 days from planting to harvesting and the interval between harvesting from the first and second cycles, that is, between harvesting of the mother plant and daughter plant, of 224 days (SILVA et al., 2013).In the present study, the first harvest of 'Prata-Anã' cultivar was at 268 days after planting, that is, 79 days earlier than in Juazeiro-BA, while the interval between harvest of the first and second cycles was similar in both sites, with a slight difference of six days.
The period between flowering and harvesting of banana plants evaluated varied with production cycles (Table 5) and with cultivars (Table 6) independently.For the variation of this characteristic among cycles, it was observed that the shortest period between flowering and harvest, 125 days, was recorded in the second cycle.The variation among cultivars was small, with the formation of only two groups, and the longest periods between flowering and harvesting were observed for 'Prata' cultivar.
In general, considering greater production and resistance to diseases, 'Maravilha' and 'FHIA-23' cultivars are an alternative for planting in conductive environments for Panama disease; referring to 'Prata'-type cultivars, BRS Platina stands out with characteristics closer to those of the genitor; for cultivars with 'Maçã'-type fruits, the hybrids, for being tolerant to Panama disease.Among banana plants with 'Pacovan'-type fruits, the most productive is 'PV79-34'; however, it is susceptible to Panama disease (RODRIGUES FILHO et al., 2014), as verified in the present study.
In addition to performance and disease resistance, in order to recommend a cultivar, it is necessary to consider commercial acceptance to minimize the risk of low adoption or rejection by producers.

CONCLUSIONS
'JV42-235', 'Japira' and 'Pacovan-Ken' cultivars are larger in size, and 'Grande Naine' is smaller.'Prata-Anã' cultivar has higher number of leaves at harvest and leaf area index similar to the others.'BRS Platina' cultivar is the earliest in flowering and harvesting.

TABLE 3 -
Yield of banana cultivars in three production cycles.Guanambi-BA, 2010-2013.Means followed by equal letters, lowercase in rows, do not differ from each other, by the Tukey test (p <0.05).Means followed by equal letters, upper case in columns, belong to the same group by the Scott-Knott criterion (p <0.05).

TABLE 6 -
Number of days between flowering and harvesting of banana cultivars in three production cycles.Guanambi-BA, 2010-2013.Means followed by equal letters, upper case in columns, belong to the same group, by the Scott-Knott criterion (p <0.05).