Background

In 2018, around 155 million metric tons of banana were produced around the world, of which 27% came from sub-Saharan Africa (SSA) (FAOSTAT 2020). The majority of this production comes from small plots and backyard gardens. The highest per capita consumption of banana in the world is in the East African highlands, where one-third of the people depend on this crop as a staple food—the crop occupies between 20 to 30% of the acreage under cultivation (Karamura et al. 2012). In Uganda, millions of people rely on banana for income and daily food, with approximately 75% of farmers cultivating banana (Jogo et al. 2013; Ochola et al. 2013).

Over the past decades, new banana cultivars have been introduced across SSA to alleviate declining yields, contribute to household food security, and improve livelihoods (AATF 2009; Aïtchédji et al. 2010; Dzomeku et al. 2007; Gaidashova et al. 2008; Lemchi et al. 2005a, b; Nowakunda et al. 2015; Ortiz et al. 1997; Pedersen 2012; Swennen et al. 2000; Uazire et al. 2008). Adoption rates of introduced banana cultivars are often low compared to their economic importance, and rates are lower than those of other staple crops (ISPC, SPIA 2014; Ortiz 2011; Walker and Alwang 2015). Studies that report adoption rates for new banana cultivars in SSA are scarce (De Weerdt 2003; Faturoti et al. 2006, 2009; Kagezi et al. 2012; Nkuba 2007). Reasons given by farmers for low uptake include inferior taste, poor marketability compared to local cultivars, and risks associated with growing new cultivars (Kagezi et al. 2012). Farmers indicate preference for local cultivars because of their superior consumption attributes (good taste, soft food texture, good aroma, and good/yellow color), even if new cultivars have better agronomic traits and better response to biotic and abiotic stresses (Akankwasa et al. 2013b; Barekye et al. 2013; Nwachukwu and Egwu 2008).

Understanding trait preferences of farmers, consumers, and other value chain actors is a first step for developing a demand-driven breeding program. Developing new cultivars and their subsequent dissemination and adoption is a complex process that starts with setting breeding objectives and developing a selection strategy for priority traits. Such a consultative process requires open dialog and collaboration between plant breeders, other researchers including social scientists, farmers, and other users such as traders and consumers, to understand the needs and preferences of different users, traits and their importance (Christinck et al. 2005).

Collecting trait information according to the role and position that an actor occupies in the value chain, as well as gender-specific information, yields wide-ranging and relevant knowledge about cultivars, their traits, and specific uses. The needs and preferences of men and women end-users intersect with various socio-economic and cultural factors at the individual (e.g., age, marital status), household (e.g., wealth), and community (e.g., culture, ethnicity) levels. These factors affect the adoption of new banana cultivars. Farmer preferences may not be the same as market traders and consumers (Ferris et al. 1997). Knowledge of traits that various end-users prefer will enable researchers and farmers to produce marketable cultivars with acceptable attributes (Mugisha et al. 2008).

The objective of this review article is to identify trait preferences reported by farmers and other actors in the banana value chain in SSA. We privilege gender-specific differences in trait preferences and the extent to which preferences can set breeding priorities in order to focus on the importance of gendered knowledge in improving food security and banana-based livelihoods. Results will be discussed with the objective to inform future banana breeding research on trait preferences that consider gender-specific needs, while developing product profiles for new cultivars.

Methods

We accessed English-language publications in both scholarly and gray literature from Musalit (www.musalit.org, repository of references on banana) and CG Space databases, using search terms that include: banana, attribute, trait, gender, preference, choice, priority(ies), end use, desirable, improved variety(ies); refer to Electronic Supplementary Material (ESM) File 1: Table S1 for a full list of the search terms. The main inclusion criteria were that the publication identified and documented banana trait preferences or cultivar preferences by end-users.

The initial screening filtered articles based on a review of their titles and abstracts using the inclusion criteria and generated 3489 articles (including duplicates). After the first round of screening, irrelevant articles were excluded. The remaining 86 research articles were screened again with a full-text reading. We then used reference snowballing to identify additional articles that the original search had missed. We excluded articles that did not meet the criteria at full document reading.

We present end-users’ trait preferences according to the specified “trait” as well as the “trait state.” Trait refers to a feature, attribute, or quantifiable measurement that can be described (e.g., taste, bunch size), while trait state refers to the observed or experienced state of the trait (e.g., sweet taste, big bunch). For breeders, “trait refers to a genetically determined characteristic that is associated to a specific phenotype.” The phenotype is controlled by its genotype (G), the environment (E) where the plant grows, and the G × E interaction (Bechoff et al. 2018, 8–9).

Results

Results presented below are based on a full, analytical review of 44 articles published between 1994 and 2018. The reviews represent ten country experiences (Table 1), the majority from Eastern Africa (70%). Overall, 45% of the articles were from Uganda. Results were differentiated according to the four banana uses common in SSA—cooking, beverage/beer, dessert, and plantain. Karamura et al. (2012) and Swennen and Vuylsteke (1991) provide detailed descriptions of these types.

Table 1 Geographic scope of studies by banana type.
Full size table

Key information vis-á-vis the geographic location of the study, data collection method, type of banana being studied, end-user’s banana trait preferences specified in the studies is presented in ESM (Electronic Supplementary Material) File 1: Table S2.

Five main categories of banana traits emerged from the studies, and drawing from (Ssemwanga 1995), include: 1) physical traits (agronomic, size, and shape attributes); 2) sensory/organoleptic traits (texture, flavor, appearance attributes); 3) processing and product-related traits; 4) socio-economic and cultural traits (cultural, commercial, and market life attributes), and 5) other traits.

End-users are likely to prioritize different traits depending on factors that may include: role in the value chain (that may be gendered), end use of the crop (determined by cultivar characteristics), environmental constraints, geographic location, individual and household characteristics, and cultural factors. The list of traits is long, making prioritization for breeders challenging.

Using a summary of preferred attributes, we grouped banana traits into the five above-mentioned categories (Table 2). When available, we provide country-specific details or nuance to the specified traits in the corresponding table narrative. Several of the studies document end-user’s preference in order of importance or highlight priority traits (Table 3, General Ranking of Banana Cultivars Irrespective of Type section). A discussion on the rankings and classification of traits’ importance is provided for each banana type (if a study exists), providing breeders with additional information on mentioned traits (Cooking Bananas, Beverage/Beer Bananas, Dessert Bananas, Plantains sections). For all banana types, end-users mention common preferred traits linked to production constraints, particularly host plant resistance to pests and diseases, high yield to ensure food security and surplus production, high market demand, and price.

Table 2 Preferred traits for each banana type identified in the review.
Full size table
Table 3 Ranking of end-user preferred traits in bananas.
Full size table

Farmers’ Trait Preferences for Banana

In their roles as producers, processors, marketers, and consumers of banana, farmers and farming households prefer a large range of traits.

Cooking Bananas

There are regional differences in preferred texture for cooking bananas; for example, farmers in Uganda prefer soft matooke cultivars (Akankwasa et al. 2013b, 2016; Barekye et al. 2013; Nowakunda et al. 2000; Nowakunda and Tushemereirwe 2004; Rutherford and Gowen 2003). In some parts of Tanzania, cultivars with a hard texture are preferred (Kibura et al. 2010). Characteristics include post-harvest attributes related to processing and value addition. Farmers prefer multi-purpose cooking cultivars that also produce juice and beer (Gaidashova et al. 2005; Nkuba 2007; Rutherford and Gowen 2003). Women value the cultural importance of banana in birthing ceremonies and food preparation, while men emphasize their use at funerals (Musimbi 2007). In one Ugandan study, women indicated that they preferred their traditional cooking cultivar “Katetema” because of its cultural values (Musimbi 2007). Farmers mention preference for cultivars that ensure normal sugar levels after eating them (Dzomeku et al. 2008). Consumption traits, such as good food quality, good taste, soft food, and good flavor, ranked high in Uganda (Akankwasa et al. 2013a, b; Barekye et al. 2013; Nasirumbi et al. 2018; Ssali et al. 2010).

Beverage/Beer Bananas

Beverage bananas are used in the production of juice, local beers, and local gin. Trait preferences are related more to the products rather than the plant itself or its fruits. Astringency, a characteristic of East African highland banana (EAHB) beer cultivars, is preferred for beverage and medicine production (Karamura et al. 2004). Farmers prefer cultivars that can be continuously de-leafed to provide leaves for steaming food, wrapping, and for sale without damaging the cultivar (Rubaihayo 1991) as well as cultivars that produce palatable food in times of food shortages (Musimbi 2007; Rubaihayo 1991).

Dessert Bananas

Organoleptic and market related attributes are key since dessert bananas are eaten raw and often sold (Ayinde et al. 2010; Kibura et al. 2010; Kwach et al. 2000; Mugisha et al. 2008; Ocimati et al. 2014; Uazire et al. 2008). Dessert bananas are often used for producing beverages (juice and wine) and snacks, hence characteristics related to the quality of the processed products are mentioned.

Plantains

Plantains are typically processed through boiling, roasting, deep frying and pounding to make food, chips, flour, and biscuits among others (Ekesa et al. 2012; Ubi et al. 2016). Traits related to the appearance of the fruit before processing and product attributes are mentioned. The pulp of “Apem” (small-fruit French plantain) is favored for a dish called Ampesi (where pulp segments are boiled until soft) as it is crispier, firmer, tastes better, and gives the best mouth feel compared to other cultivars in Ghana (Dadzie and Wainwright 1995). In Cameroon, farmers ranked attributes in order of importance as follows: bunch size, fruit length/weight, taste/softness of pulp, and early maturity (Mengue Efanden et al. 2003). Long banana mat perpetuation is preferred (Mengue Efanden et al. 2003). Rwandese farmers reported early maturity as an important criterion (Ocimati et al. 2014, 2016).

General Ranking of Banana Cultivars Irrespective of Type

Differences in trait rankings based on the geographical location and type of banana grown exist (Table 3). Gold et al. (2002) showed regional differences in the relative importance of banana cultivar selection criteria in Uganda. Principal component analysis (PCA) revealed that farmers preferred drought tolerance, marginal soil tolerance, and longevity (as determined by the first principal component or PC1), which means a robust cultivar that grows as a perennial but with fewer inputs (i.e., a labor-saving cultivar). In the second principal component (PC2), ripening and post-harvest characteristics were preferred (bunch size, taste, maturation, marketability). Dowiya et al. (2009) found regional differences in selecting banana planting material types in North and South Kivu of DRC where plantains and beer bananas are popular. In South Kivu, the most important criterion was a combination of flavor, taste, and juice production; in North Kivu, it was pest resistance.

Farmers’ selection criteria also reflect the major challenges faced in banana production. In areas where soil fertility is low, or where incidences of pests and diseases are high, adaptability to low soil fertility and resistance to pests and diseases would be critical selection criteria (Ocimati et al. 2016). Following outbreaks of Xanthomonas wilt, farmers might switch to other crops if alternative banana cultivars with the traits they prefer are not available; e.g., in Uganda farmers switched to sweet potatoes and cassava to cope with the devastation of fields by Xanthomonas wilt (Kalyebara et al. 2006; Karamura et al. 2006).

Trait Preferences of Other Actors in the Banana Value Chain (Outside the Farming Household)

Consumers and traders have their own preferred trait preferences. For cooking bananas and plantains, consumer trait preferences are determined by the product type and processing method (Dadzie and Wainwright 1995; Dury et al. 2002; Dzomeku et al. 2006, 2008). Consumers indicate preference for cultivars whose fruit are firm and crunchy when boiled and soft for fufu preparation. At the time of purchase, consumers prefer cooking bananas with big bunches and big fruits that are fresh. Price and the type of biotechnology used to produce the planting material are taken into consideration (Kikulwe et al. 2011; Nalunga et al. 2015; Pillay and Tenkouano 2011). With respect to dessert bananas, consumers prefer yellow skin color, light yellow pulp color, no spots on peel, soft, sweet, firm fruits with easily separable skin and easily detachable fruit. Urban consumers in Nigeria purchase dessert banana fruits according to taste, fruit size, number of fruit/hand (cluster of fruit), texture, aroma, and shelf life of 9 to 12 days (Ayinde et al. 2010). With respect to plantain, fruit shape, fruit size, aspect of the fruit, ripening/maturity stage, bunch size, and good textural qualities after cooking (depends on product and processing method) are desired attributes (Dadzie and Wainwright 1995; Dury et al. 2002; Dzomeku et al. 2006, 2008; Kouamé et al. 2015). Kouamé et al. (2015) found that for urban consumers in Cote d’Ivoire, plantain ripening/maturity stage used to prepare different foods was more important than other physical attributes.

Traders prefer cooking bananas with big bunch sizes, long fruit, more fruit per bunch and per hand, big hands, more hands per bunch and compact bunches for easy transportation. With respect to appearance, traders prefer cooking cultivars with a good fruit sheen and a pale green fruit color. Other commercial and market life attributes for cooking banana include hands and fruits that do not easily fall off (fruit drop), gradual ripening down the bunch, no easy bruising or quick wilting, a long shelf life, low rate of sheen loss, and price (Akankwasa et al. 2013b; Nalunga et al. 2015; Ssemwanga 1995). For plantain, traders prefer large bunch size (Dadzie and Wainwright 1995; Dury et al. 2002; Dzomeku et al. 2006, 2008; Kouamé et al. 2015).

Gender-Differentiated Banana Trait Preferences

Literature on gender-differentiated trait preferences in the banana value chain is scarce. Although several studies focus on end-user’s banana trait preferences (N = 44), only four of these provided some form of gender-specific data (Edmeades et al. 2004; Miriti 2013; Musimbi 2007; Nasirumbi et al. 2018). Table 4 presents the review’s gender-specific trait preference findings. These studies documented farmers’ preferred traits for cooking and dessert bananas but did not examine trait preferences directly, except Nasirumbi et al. (2018). Rather the focus was on household cultivar demand, gender-responsive strategies in banana production, and impact of gender on adoption.

Table 4 Trait-preferences according to respondents’ gender.
Full size table

Statistical differences in the importance of banana traits between men and women banana farmers were found for cooking quality (taste, color, softness), beer quality, and resistance to Fusarium wilt. Differences were attributed to underlying preferences based on gender roles; men /beer production and women /cooking, respectively (Edmeades et al. 2004). Miriti (2013) provided male and female farmers’ preference rankings for banana cultivars, but did not specify traits they preferred in the different cultivars.

Discussion

This article contributes to general knowledge of banana trait preferences, including gender. It illuminates the needs and preferences of farmers and banana value chain actors that can be used to orient “product profiles” for new banana cultivars of the different banana types in different ecologies, recognizing the significance of gender-specific trait preferences (Weltzien et al. 2020).

Product profiles are used for priority setting for breeding cultivars of matooke and mchare cooking bananas, types popular in Uganda and Tanzania, and are publicly available (http://breedingbetterbananas.org/wp-content/uploads/2018/07). At the time of submitting this manuscript, the product profile for plantain was not yet publicly available (pers. comm. R. Swennen, November 2019; rony.swennen@kuleuven.be). There is no published product profile for beer or dessert bananas. The published banana product profiles include some of the traits identified above and could be expanded. Product profiles mostly include production and adaptation related traits, such as pest and disease resistance, suckering ability, early maturity, tolerance to drought, and resistance to wind (through plant height). Traits not currently included in profiles include: agronomic attributes (e.g., adaptability to poor soils); processing traits related to value addition (e.g., size and shape attributes, such as uniform fruit size, straight fruit for ease of peeling, and compact bunches for easy transport); social and cultural traits—plant parts which can be used for multiple purposes (e.g., banana leaves for use in food preparation or roots for medicines). End-users mention contrasting traits such as big bunches for the market and small bunches for home consumption.

Although sensory/organoleptic/consumption traits are included in product profiles, they are categorized under one umbrella, and treated as a single trait: “table quality/palatability.” There might be need to separate the traits in this category, as color, taste, or texture are highly complex characteristics. For example, traits such as “good textural quality after cooking and suitability for various uses” or “firm and crunchy when boiled and soft for fufu preparation” indicate specific demands from consumers, which could need to be incorporated into the product profile.

Such consumption and processing attributes are poorly understood in terms of assessment (measurement), inheritance, and their physicochemical nature. Physicochemical characterization, molecular assessments, and interdisciplinary work with food scientists and geneticists would increase the options for inclusion of such traits.

The relative priority of different traits in new cultivar design is an important process in breeding. The review presents a long list of traits for each banana type, requiring trait prioritization to set breeding goals and objectives. It is not feasible to include all trait preferences in a banana breeding program due to limited resources and time. A decision-tree analysis, with the critical actors in the value, is one way to prioritize traits and address conflicting factors (Shimelis 2017).

Banana Trait Preferences

Different banana types share common preferred traits linked to production constraints, such as resistance to pests and diseases, high yield, and high market demand/prices. The review found more preference studies for cooking and plantain types than for dessert and beer banana types. Cooking and plantain types share several common traits, especially related to processing and consumption (appearance, texture, and flavor attributes) as “cooking” types. Dessert and beer bananas are processed into juice and other beverages that may contribute to household income, hence traits related to yield, flavor, taste, and quality of beverage products are mentioned. Traits like fruit length appear to be less important for marketing as each size has its own market, indicating these traits have a wide range of acceptable states.

This review found that traits such as host plant resistance (e.g., black leaf streak, Fusarium wilt, and weevils), abiotic stress tolerance (e.g., short plants and strong root systems to avoid wind damage), superior agronomic performance (heavy bunch with big fruit sizes), and vegetative propagation (as related to suckering behavior) are traits that farmers mention and prefer, and are those that breeders target in their programs (Brown et al. 2017). While there appear to be some common priority attributes, cultivars are more likely to be selected if they are better adapted to a region’s agro-climatic conditions, local farming systems, and show resistance to prevailing pests and pathogens.

Superior consumption attributes, such as taste, flavor, pulp color, and other fruit post-harvest traits (e.g., pulp texture, shelf-life) appear to be of critical importance for cultivar preference and thus adoption of new banana cultivars. Farmers in different regions, however, prefer different banana types for different end uses and cultural events, and hence prioritize different consumption and use related traits. For example, farmers in some regions in Tanzania prefer cooking banana types with a hard texture (Kibura et al. 2010; Pedersen 2012) like mchare, whereas in Uganda consumers prefer EAHB cooking cultivars that make soft food. Edmeades et al. (2004) and Tenkouano et al. (2010)  argue that ethnicity strongly influences some of the preferences for organoleptic attributes, such as taste, color, and feel of food.

A small number of studies focused on trait preferences of other actors in the banana value chain indicating that cultivars should be marketable (high market demand) and have traits that other value chain actors (traders, processors, consumers) prefer. Farmers sell surplus to consumers and traders in local, urban, or regional markets where preferences are likely to differ by region.

Consumers who are not producers are versatile—their consumption patterns depend on what is available in the market. They may substitute or switch to a different product (e.g., rice or potato) if the preferred banana cultivar is not available. Additionally, household demand from urban consumers for a certain product depends on income level.

In view of this high diversity of demands for a wide range of traits, banana breeding programs need reliable, detailed information about agronomic, use, and market-related trait preferences of their potential customers. This information can help to identify traits and trait complexes that are important for a large proportion of priority customers. The breeding programs can target improvements for such priority traits, and thus improve the chances that the new cultivar can be beneficial for a large proportion of farmers, farm families, and possibly other consumers. In addition, there is a need to holistically understand the traits and the diverse factors that may affect preferences and eventually other factors that influence adoption. Thus, the revisions of product profiles need to reflect the available understanding of consumer demands in terms of trait combinations and acceptable trade-offs, including gender impacts. As available evidence and trends that could lead to changes are scarce, breeding programs would benefit from well targeted forward-looking consumer studies.

Gender-Differentiated Trait Preferences

The review found that male and female banana farmers often have similar production constraints or common goals such as food security or ceremonial uses, and that men and women might prefer cultivars with big bunches and fruit with a commercial value. Musimbi (2007) found that women mentioned traits related to production (high suckering ability and early maturity), whereas men emphasized consumption-related traits (good taste and color). Women preferred high-suckering cultivars given the potential to earn higher income from selling suckers.

We contend that potential differences in preferences, which are not specifically discussed in the reviewed studies, might stem from the different roles that men and women play in the banana value chain, e.g., cooking attributes for women and beer production for men (Edmeades et al. 2004). Women are traditionally responsible for food preparation and processing of banana (Musimbi 2007), whereas men are involved in the preparation of juice that can be fermented to produce local beer (Edmeades et al. 2004; Musimbi 2007; Nkuba 2007). Studies done in Kenya and eastern Uganda indicate that women predominantly participate in ripening and marketing activities at local markets, along roadsides, in trading centers, and in nearby schools (Miriti 2013; Musimbi 2007), whereas men sell at organized markets or farm gates which might lead to differences in preferences. Differences in production goals may lead to varied preferences. Men and women might also face different constraints e.g., related to mobility, information, and input constraints that may affect adoption of new cultivars (Christinck et al. 2017; De Weerdt 2003; Musimbi 2007). Such constraints need to be addressed when designing breeding programs.

Farmers often associate new cultivars with increased labor burdens. Musimbi (2007) notes that the introduction of FHIA banana cultivars required digging bigger holes, use of more crop residue, farmyard, and animal manure, and de-leafing in order to produce big fruit. This increased work burden might have different negative implications for men and women farmers, depending on their roles and responsibilities in the production system.

Gender roles, constraints, opportunities, and preferences are not static. Information on trait preferences, though, is often collected at only one point in time and may not include on-going changes in gender relations. Recognizing such change will support the targeting of breeding programs towards needs of their priority customers taking socio-economic and agronomic factors, such as geographic location, gender, ethnicity, culture, age and their interactions into account. Overall, the few studies focusing on gender-specific information indicate that it is essential to capture gender-differentiated preferences of actors in the whole value chain to improve the chances that new cultivars can be adopted and generate maximum benefit.

Conclusion

The finding that farmers often prefer traditional cultivars because of their superior consumption attributes, even if new cultivars have better agronomic and host resistance characteristics, is a recurring theme of the reviewed studies. Using local germplasm to produce new cultivars can potentially improve acceptance rates, especially as these cultivars would meet the farmers’ and consumers’ preferences for taste, color, and processing related traits. Understanding what end-users and farmers want in cultivars early on can assist breeders with appropriate targeting of efforts. Bridging the divide between farmers and breeders is one way to ensure that new cultivars have farmers’ desired traits, which might lead to faster adoption.

Sustained interaction between breeders, other researchers such as pathologists, agronomists, food scientists, social scientists, and entomologists, farmers, and other value chain actors is necessary to understand local context and exchange vital information for an efficient and effective breeding program. Interdisciplinary teams can build “product profiles” for improved cultivar banana types that may be highly acceptable to well-targeted farmers in specific prioritized growing regions (Ragot et al. 2018). Preference studies provide entry points for discussions that prioritize targets for the improvement of specific traits, and priorities for selection in the short and longer term. This can contribute significantly to enhancing the efficiency of a breeding program by improving the chances that the new cultivar would be adopted by farmers and contribute to improving livelihoods. A cultivar product profile based on the priority needs and preferences of priority end-users can be the basis for developing an effective breeding strategy.

The reviewed publications contribute only partial information to building such profiles. Research documenting successes and failures of past cultivar releases, adoption rates of introduced banana cultivars, adoption rationales, and a better understanding of the farming production and seed systems remains scanty or missing. Collecting information and understanding why some cultivars are more popular than others is recommended. Popular cultivars are more likely to have traits that end-users prefer, which can help guide breeding programs on what traits to target. Banana breeders need quantifiable information on trait preferences and guidance to set priorities for selection. Traits ranked in order of importance by end-users can provide useful information to help banana breeding teams to adapt and revisit the product profiles and breeding priorities (Ragot et al. 2018).

Finally, the review did not find research that evaluated gender differentiation from a value chain perspective. However, as men, women, and other social groups such as traders might have gender-specific knowledge on production, processing, or consumption of particular cultivars, a gender approach can improve efficiency of the breeding program by contributing to the development and refinement of breeding product profiles (Christinck et al. 2017; Ragot et al. 2018).