Globally banana (Musaspp.), is a multipurpose tropical plant produced mainly for human nutrition, health benefits and industrial processing [1]-[3]. In Sub-Saharan Africa, Musa banana is also a major livestock feed [4], soil fertilizer [5][6], like elsewhere, a resource for building shelters in rural areas [7][8] and making hand crafts [9].

A perennial crop that reproduces through overlapping cycles, with a yielding potential of 100 t/ha, a banana plant has a very high soil water utilization rate of ≥5 litres per day during early plant establishment and ≥25 litres per day for reproductive plants. Banana macro-nutrient (Nitrogen-N, Potassium-K, Phosphorus-P, Calcium-Ca, Magnesium-Mg and Sulphur-S) and micro-nutrient (Boron-B, Copper-Cu, Iron-Fe, Manganese-Mn and Zinc-Zn) requirement for optimal growth, fruit development, carbohydrate synthesis and translocation in its above and below ground tissues (leaves, stems, fruits, corms and roots) have been documented [10]. A balanced amount of both macro- and micro-nutrients supplemented with sufficient water supply all year around ensures enhanced banana productivity [11]. The key mineral nutrients required by banana include potassium, nitrogen, magnesium, sulphur, boron, copper, iron, manganese, molybdenum, and zinc, but the level of consumption of these nutrients varies with cultivar, phenology and environment [10], which significantly affect banana palatability [12]. Because of its high nutrient demand, various recommendations have been made of which application of 7 - 8 Kg N, 0.7 - 1.5 Kg P and 17 - 20 Kg K per metric tonne per yield was found ideal, and based on a study carried out in Uganda, a single blanket rate of 100N-30P-100K-25Mg kg per hectare per year was recommended [11].

The East Africa Highland Banana cooking type exists in numerous varieties of which more than 50 cultivars are grown in the South Western Agro-Ecological Zone of Uganda. Marketability or consumer preference influences the level of variety adoption by farmers, where varieties that are high yielding, produce big bunches, highly abiotic stress tolerant and possess preferred sensory qualities are priority. Also, for dessert banana/plantain Musa species (spp.), farmers and end users prefer varieties with appealing and good agronomic attributes. Farmers’ and end-users’ most desired traits are high yields, good cooking qualities [13] and climate resilience [14]. Previously, Mulugo et al., revealed that gender influences banana cultivar adoption, where men preferred high-yielding varieties with desirable traits like bunch and finger size, while women preferred early-maturing cultivars with specific organoleptic banana food attributes: pulp colour, food colour, texture, aroma and taste [15]. Even if new cultivars have better agronomic traits and better response to biotic and abiotic stresses, research has shown that farmers still prefer local cultivars because of their superior consumption attributes (good taste, soft food texture, good aroma, and good/yellow colour) [16]-[18].

Out the numerous (~130) banana varieties grown in Uganda among the banana growers and consumers in both rural and urban areas, the local cultivars and landraces are the most common [19][20], and preferred compared to the improved cultivars. Even when the improved cultivars have been developed to address major problems associated with agronomic traits, abiotic and biotic stresses, the rate of improved or hybrid banana adoption by farmers has remained very low in the South Western Uganda [21][22]. In a study to assess farmers’ choice among developed hybrid banana varieties in Uganda, the local Mbwazirume, was the most preferred by farmers because it had large bunch sizes, good food taste, soft texture and good flavour. While for years, banana improvement research has been focused on developing varieties with superior agronomic traits, a study found out that quality characteristics were rated higher than the agronomic performance among the farmers, the market (traders) and end users/consumers. In the above study, five (Nakitembe, Mbwazirume, Kibuzi, Musakala and Mpologoma) of the following varieties (Nakitembe/Entaragaza, Kibuzi, Enyeru, Mbwazirume, Musakala, Mpologoma, Enjagata, Muvubo, Kisansa, Embururu, Rwamigongo, Nakyetengu, Majaga, Butobe, Enzirabushera, Entukura, FHIA, Kabana, Lusumba, Nakawere and Nakinyika), were identified for making good banana food as the most preferred because of their softness, taste, good yellow colour, big finger size and big bunch size [21].

The vulnerability of the local banana cultivars to climate change, which continues to negatively affect the plant phenology and productivity due to the changing unfavourable rainfall patterns and higher temperatures, negatively impact banana production and farmers’ returns to investment in Uganda. Under climate change, increasing unfavourable weather characterised by prolonged dry spells and erratic rains leading to water stress have become common and create favourable conditions for increased pest and disease prevalences, which further threatens banana cultivation by escalating crop damage and yield losses. These impacts create significant challenges for food security and farmer livelihoods, necessitating adaptation strategies such as developing abiotic and biotic-stress resistant cultivars [23].

Cooking hybrid varieties were developed to improve banana productivity and production in Uganda. With evidence, the hybrids are high yielding, and produce big banana bunches even under severely unfavourable conditions under which the local cultivars totally fail [13][16]. Hybrids produce and maintain many green leaves that are disease-free, which support development of big fruit bunches. However, limited research attention has been paid to the quality attributes based on hybrid banana palatability/sensory analysis by consumers in the South Western Uganda. The primary aim of this research was to analyse the consumers’ acceptability of new cooking banana hybrids in comparison with local cultivars/landraces, utilizing matrix scores and preference ranking based on the four key culinary attributes. The study evaluated the palatability of East African Highland Banana cooking varieties (hybrid and local cultivars/landraces) through sensory evaluation of unpeeled steamed fingers (Empogora), peeled steamed-mashed pulps (Emibumba commonly known as matookein Uganda) and fresh juices, using the matrix scores and logistic preference ranking analysis [24].

The palatability analysis and preference ranking were carried out using seven hybrid banana lines (NARITA 4, NARITA 17, NARITA 18, NARITA 22, NARITA 24, NAROBAN 4/NARITA 27 and NAROBAN 5/NARITA 30), described in Table 1. These varieties were developed by the National Agricultural Research Organization (NARO) in collaboration with the International Institute of Tropical Agriculture (IITA). The tissue culture plantlets were provided by Dr. Ivan Kabiita Arinaitwe a Plant breeder formally based at the Banana Programme of the National Agricultural Research Laboratories Kawanda of the National Agricultural Research Organisation, Kampala, Uganda. Ninety-five tissue culture plantlets of the improved banana hybrids and 15 plantlets of a local variety (Enzirabahima) were planted on the 12^th December 2022 in a randomized complete block designed experiment, with eight 21 m × 12 m plots replicated three times in the field at the Mbarara Zonal Agricultural Research and Development Institute (MbaZARDI). For sensory evaluation, three fully mature bunches were harvested from each of the five plants of the seven hybrids, in each replication. The test samples of the two desserts (FHIA 25 and Apple banana/Kabaragara) and the local market preferred cooking cultivar (Mbwazirume) (Table 2) were similarly collected from a similar trial that was established for the non-hybrid types in the same period, environment and grown under the same agronomic management system and cycle. Three fully mature banana fruit bunches were harvested from each of the five mother plants per cultivar per replication, and the middle fingers from the

Table 1. East African highland banana cooking hybrid, dessert and the local cultivars tested for palatability in the South Western Agro-Ecological zone of Uganda.

Sources: Tushemereirwe et al., 2015 [40]; https://www.promusa.org/NARITA+hybrids [41]. Karamura et al., 2012 [20]; Ilukor et al., 2025 [42].

Table 2. Phenotypic characteristics of the East African highland Banana cooking hybrid, dessert and the local cultivars grown for palatability evaluation in the South-Western Agro-Ecological Zone, 2022-2024.

second cluster from the bunch-top were prepared for sensory testing as unpeeled steamed fingers, peeled steamed-mashed pulps and fresh juices.

Cooked banana food was prepared using 10 fingers that were uniformly picked from the middle of every second hand of a fully mature bunch of the above nine cooking test cultivars. Fresh unpeeled and peeled banana fingers were washed and steamed in banana leaves until they were well-cooked and ready to eat. The well-cooked steamed peeled banana pulp of each variety was gently mashed until it became uniform and smooth for the further steaming on low heat to get it softer and flavoured. The unpeeled cooked fingers were maintained on lower to keep them hot and soft.

For juice preparation, well matured bunches of the 11 cultivars that were ripened to stage five (peel colour is bright yellow, with minor green typically found only at the very tips of fruit, fruit is still relatively firm with pleasant sweetness and a good texture, a perfect ripeness for eating with a sweet flavour) [25], under natural conditions were separately peeled. The pulp of each cultivar was added to corresponding containers of hot water at the temperature of 100˚C in a ratio of one kilogram (Kg) of banana pulp to two litres of water [26]. After adding the banana to the hot water and mixing, the temperature dropped. The mixtures were reheated to restore and maintain their temperatures at 85˚C for 30 minutes. After 30 minutes, each mixture was strained or filtered into clean containers using a cheesecloth secured with a strong string. The resulting fresh banana juices were further cooled and maintained at room temperature, awaiting evaluation.

The banana food samples of steamed whole fingers and mashed pulps were tested and evaluated by a 16-member panel, while the two-juice cycles were tested by a similar 10-member panel. To be selected as a panellist, individuals that had banana as their major food; liked/enjoyed eating the banana as cooked food, ripe fruits and juices; were adults who were interested in participating and voluntarily consented by registering/signing were allowed to participate. The selected panellists were trained on how to use their four senses to test the colours, flavours, textures and tastes of different banana variety food samples based on the 1 - 5 matrix scales and preference ranking. On day 1, the two hot steamed banana samples were served, and well labelled using codes for the 16 panellists to assess them without bias, and take notes as required in the data collection sheets. Each panellist tested all banana samples as steamed-fingers and steamed-mashed pulps. A glass of clean drinking was provided to each panellist to be used to rinse his/her mouth between the different sample tests. The freshly made juice samples of 11 varieties were evaluated by 10 panellists in two cycles. Because different banana varieties tend to have varying maturity periods especially when the plantation is still young and the mother plants have not produced many suckers, we failed to harvest mature banana fruits from all the varieties for the first cycle, so, every set of banana varieties that would have mature fruits would be tested. Therefore, first cycle combined results from the 11 varieties separately tested. In the second cycle, all 11 varieties had mature fruits for us to uniformly extract fresh juices as described in the first cycle. Each well-labelled set of 11 samples of juice was randomly served, and analysed by the different individuals in the panel at once.

Brix or total soluble solids (TSS), which is primarily sugar content a direct indicator of the sweetness and flavour of banana juice was measured using A Contrast 0% - 32% Brix ATC Refractometer or ERMA Hand Refractometer (ERMA INC. Tokyo Japan), which is calibrated to measure 0 - 32 g per 100 g of the sample (2 - 3 drops). This was done to find out the cultivar that produced juices that meet the consumer’s preferred taste, since high Brix level typically correlates to a richer and sweeter flavour. A few drops of each extracted, filtered, non-diluted and diluted juice were placed on the prism of a hand-held refractometer that had been calibrated with distilled water [27]. The refractive index of each sample juice was converted by by the refractometer to a percentage Brix value and recorded.

Palatability data of unpeeled steamed fingers, peeled steamed-mashed pulps and fresh juices, were collected based on sensory parameters that included: colour, flavour, texture and taste using the matrix scale of 1 - 5 (1 = very bad, 2 = bad, 3 = fair, 4 = good, 5 = very good). The 1 - 5 scale was used in the sensory evaluation because it correctly modelled human sensory qualitative ratings, which were categorical and ordered as quantitative variables (1 = very bad to 5 = very good) in analysis; it was suitable for scoring the uniformly prepared banana variety samples across the four test parameters; it gave specific numeric figures for every response; its narrow range minimised the range of variation or error within the data sets, and it was easy to apply by the selected individuals. The varying data generated by the different members of the panels, to determine whether the recorded sensory data were actual, and not merely results of uncontrolled variation in responses, the One-Way ANOVA was carried out to assess the statistical significance of variations within each set of data per attribute/parameter. The analysis based on one-way ANOVA (analysis of variance) was conducted to compare the effect of the banana variety [independent variable] on sensory attributes [dependent variable] using the single categorical factor, while effectively preventing errors. The matrix scores (1 - 5) and preference ranking data sets were entered in MS-Excel and analysed using the Statistix analytical software version 10. The means that revealed statistically significant differences among the varieties were separated using the least significant difference (LSD) at the probability level of 5% with the following formula: L S D A , B = t 0.05 / D F W M S W ( 1 / n A + 1 / n B ) . Where: t= critical value from the t-distribution table, MSW = mean square within, obtained from the results of the ANOVA test, n = number of scores used to calculate the means.

Preference ranking was based on the panel’s choice reasons using sensory parameters (colour, flavour, texture and taste). The most preferred varietal samples of steamed food or juice, were separately ranked number 1, while the least preferred variety was rank last under each attribute assessed by each panellist. Microsoft Excel 7 based logistic preference ranking analysis [24], was used to translate subjective sensory attributes of colour, texture, flavour, and taste into statistically robust probabilities of acceptance based on the number of times a particular variety was ranked in a specific position by the panellists. Ranking tests were used to measure relative preferences, by isolating how the specific attributes (colour, texture, flavour and taste) independently influence the overall acceptance or rejection of a banana variety, which verified and supported the matrix scores results, thus predicting the probability of the consumers’ acceptance of a specific banana variety. The level of significance for the logistic regression at 15% (p< 0.15) was used to retain all potentially important predictors, accounting for the wide variability in the panellists’ qualitative responses.

The East African Highland Banana (EAHB) hybrids were developed through collaborative research between the National Agricultural Research Organisation (NARO) Uganda and International Institute of Tropical Agriculture (IITA) to improve banana productivity for food security [28], and income generation especially for communities that are highly dependent on banana, and the contribution to the gross domestic product (GDP). In this study, seven hybrid varieties were selected based on their high level of pest and disease resistant, and stable high yielding potential across environment and season. Although, previous research studies elsewhere, showed that the sensory characteristics of the banana hybrids were closest to those of most popular local consumer-preferred cooking banana cultivars such as Mbwazirume in terms of pulp-food colour, flavour/aroma, taste and texture; hybrids have not been well adopted in Uganda [22][29], especially in the South Western Agro-Ecological Zone, where banana ranks first as the major staple food security and income generation crop among the majority farmers [16].

In the South Western Agro-Ecological Zone (SWAEZ), farmers adopt banana varieties based on a number of attributes, of which the key ones are: productivity, marketability and culinary associated attributes, implying that an adoptable variety must be high yielding, highly marketable and with very good palatability traits [30]. A study by Nasirumbi et al., 2023 [31], found out that hybrid banana varieties had very good productivity characteristics of high yields, big bunch sizes, big fruit fingers and many hands, while the local cultivars were rated superior based on the consumer-associated qualities. Local cultivars also referred to as traditional landraces, have been observed to be more preferred than the improved varieties because of their superior consumption attributes [32]. In the current study, the local Mbwazirume was consistently ranked the best as unpeeled steamed fingers and peeled steamed-mashed pulps by consumers based on the four culinary or sensory evaluation parameters (colour, flavour, taste and texture). In preference ranking, Mbwazirume was given the first ranks with the highest probabilities of acceptance by consumers. Previous studies have shown that consumer preference traits of any banana variety significantly affect its adoption, marketability and profitability [13][16][33].

Commercially, farmers prefer to grow banana varieties with good agronomic attributes, such as the hybrids, but they often face significant trade-offs, such as low consumer preference for their food’s taste, texture, flavour and colour, which makes them rejected and not adopted [29]. Because of the very short shelf-life of the cooking banana types after maturity [34], due to their high-water content, susceptibility to pests and diseases, and quick ripening rate [35], risks of significant post-harvest losses are high. The inferior attributes of the improved banana varieties, and in this case the hybrids, caused them to have high probabilities of rejection by farmers and consumers. Apart from the local cultivar (Mbwazirume) that had good average scores as unpeeled steamed finger pulps and peeled steamed-mashed pulps, the average scores of NAROBAN 5 were not significantly different from Mbwazirume,while the scores of NARITA 22, NARITA 18 and NARITA 17 followed closely to NAROBAN 5 (NARITA 30). A similar set of varieties had moderate to very good scores as peeled steamed-mashed pulps, but Mbwazirume and Enzirabahima were very good and good, respectively, while the sensory scores of NARITA 18, NARITA 17 and NARITA 22 were not significantly different from Enzirabahima. Interestingly, the current study results emphasise that the local Mbwazirume was the most highly scored cooking banana cultivar compared to the NARITA hybrids, which deviates from previous findings, where for example NARITA 4 and Mbwazirume, had high and moderate scores in preference as food in Mbarara, respectively [13]. This observation suggests probable genotype by environment interactive and/or physiochemical composition effects [29][36]. NARITA hybrids have been found to be superior in agronomic traits, while the local varieties such as Mbwazirume remain superior in culinary qualities (sensory traits) [37], although based on the current study, the food sensory attributes of some hybrids (NAROBAN 5/NARITA 30, NARITA 22, NARITA 17 and NARITA 18) were not significantly different from the good local cultivars.

In agreement with the matrix scores, preference ranking revealed that Mbwazirume maintained the first position as the most consumer preferred banana variety as steamed fingers and mashed pulps with the moderate to high probabilities of acceptance in the first ranking position, followed by NAROBAN 5 (NARITA 30) with a moderate probability of acceptance in the same position. In second ranking position, the two banana varieties maintained high probabilities of acceptance as steamed fingers, while as steamed-mashed, Mbwazirume, NAROBAN 5 and Enzirabahimaattained the high, moderate and moderate PoA, respectively. In position two, NARITA 17 and NARITA 22 had moderate probabilities of acceptance as steamed fingers, while NARITA 4 and Enzirabahima had low PoA. In the third position Mbwazirume, NAROBAN 5, NARITA 17 and NARITA 22 attained high PoA, NARITA 4 and Enzirabahima had moderate PoA, while the rest of the cultivars had low probabilities of acceptance as steamed fingers. As a steamed-mashed pulp, Mbwazirume maintained the moderate to high PoA, followed by Enzirabahima and NARITA 4, which had moderate PoA in positions two and three.

Apart from FHIA 25 and Apple banana (Kabaragara), which are well known to be juice and dessert cultivars, respectively, the cooking hybrids NARITA 4, NARITA 17 and NAROBAN 4/NARITA 27, had good average scores for juice sensory attributes (sweet taste and flavour) that were not significantly different from FHIA 25 and Kabaragara, implying that some of the test cooking hybrids may possess traits that make them suitable in juice, dessert and sweet product making. Based on the taste attribute, the best performing varieties were FHIA 25, Kabaragara, NARITA 4, NARITA 24 and NAROBAN 4 (NARITA 27). While the local cooking banana cultivars (Mbwazirumeand Enzirabahima) were most preferred as steamed fingers and mashed pulps, the hybrids (NARITA 4, NAROBAN 5 and NARITA 17) had moderate PoA in ranking positions one, two and three as fresh juices, but in position three, NARITA 22, NAROBAN 4, Mbwazirume and NARITA 24 had moderate PoA.

Considering the previous results that showed taste as the most highly scored attribute, the moderate ranking of hybrids (NARITA 4, NARITA 17 and NAROBAN 4) as the most preferred varieties for juices revealed that the hybrids contained higher sugar contents than the local cooking cultivars. These results align with the fundings of Nowakunda et al., in 2024, where NARITA 4 and NARITA 17, whose scores were close to the most preferred genotypes, suggesting that they have a high potential for adoption [29]. Additionally, because of their moderate, very good attributes as either food and/or juices and desserts, the hybrids have a potential of being adopted because of their multi-trait or multi-purpose characteristics, good agronomic performance (early maturity, high yield, big bunch size), high multi-stress (abiotic and biotic stress) tolerances, which characteristics might the most preferred local banana cultivars and landraces not possess. Considering the fact that consumer preference is the key determinant of what banana variety the market should supply and the farmer should produce [38], which also indirectly affects the profitability and sustainability of the banana trade, there is a need for breeders to first prioritize consumer preferred attributes when developing varietal and non-varietal technologies and banana-based products [35] that are expected to be accepted and adopted by farmers and consumers. The unique attributes possessed by most local banana cultivars keep farmers committed to growing them, even when some of them may be: inferior in the key agronomic traits, susceptible to pests and diseases and non-drought tolerant, are very important and should no longer be ignored in technology development [16][39].

The local or traditional East African Highland cooking banana cultivars were the most preferred based on matrix scores, preference ranking and probability of acceptance (PoA) based on colour, texture, taste and flavour of steamed finger pulps and steamed-mashed pulps, whereas the hybrids were most preferred as fresh juices, which were extracted through the hot water method. The second ranking varieties with moderate PoA in positions two and three as fingers after Mbwazirume were NAROBAN 5 (NARITA 30), NARITA 22 and NARITA 17, whereas as steamed-mashed pulps were Enzirabahimaand NARITA 4 in ranking positions two and three with moderate PoA. NARITA 17 and NARITA 22 had low PoA, non-significant positive intercept. As juices, the local apple banana had the highest scores as juice, but the juices of the cooking hybrids (NARITA 4, NAROBAN 5 and NARITA 17) had moderate PoA, high slopes and positive intercepts, implying that they had high probabilities of being accepted by the consumers. Apple banana, NARITA 4, NARITA 22, NARITA 24 and NAROBAN 4 (NARITA 27) had higher sugar contents (Brix) than the rest of the varieties. Overall, the hybrid varieties (NAROBAN 5, NARITA 22, NARITA 17 and NARITA 4) had high probabilities of acceptance as food and juices like the Mbwazirumeand FHIA 25 and/or Kabaragara, respectively.

The authors greatly thank Dr. Ivan Kabiita Arinaitwe, a Plant breeder formally based at the Banana Programme in the National Agricultural Research Laboratories, Kawanda of the National Agricultural Research Organisation (NARO) Kampala, Uganda, but currently working for the Kabale University, Uganda, for kindly providing the tissue culture plantlets of banana hybrid varieties and local cultivars used in this study. Special thanks to the technicians from the Banana Programme and Mbarara ZARDI that setup, managed and monitored the field experiments and participated in data collection; the panellists from Mbarara Agricultural Research and Development Institute, Mbarara University of Science and St. Mathias Mbuye Agricultural College Rakai, that participated in the banana palatability sensory testing activities.

This research was funded by the Government of Uganda through the National Agricultural Research Organization (NARO) research and development projects.