The study was carried out from November 2019 to March 2020 in four States of Southern Nigeria including Anambra, Enugu, Ondo, and Oyo (Figure 2), situated between longitudes 2°0′ and 10°5′ E and latitudes 2°5′and 11°0′ N (Iloeje et al., 1981). These sites were selected based on information retrieved from literature which indicated them as home of the species in Southern Nigeria (Ogunsola and Ilori, 2008; Nkwocha et al., 2014; Njoku et al., 2015). Anambra and Enugu, forming the Southeast region, are inhabited by Igbo, while Ondo and Oyo, forming the Southwest, are Yoruba populated. The climate of south Nigeria is dominated by the presence of three major wind currents, namely the maritime tropical (mT) air mass, the continental tropical (cT) air mass, and the equatorial easterlies (Ojo, 1977; Iloeje et al., 1981). The regions enjoy a tropical climate with two distinct seasons including the rainy season (April–October) and the dry season (November–March). The mean annual rainfall falls between 1,250 and 1,500 mm, whereas the mean monthly temperature ranges between 25.7°C (in July) and 30.2°C (in February). The soils are of sandy, loamy sandy or clay loamy, sandy clay, sandy loam, silty clay, and silty loam. Generally, the soil is deep and well drained (Adejuwon and Ekanade, 1988; Nkwunonwo et al., 2020).

The respondents were selected by combining convenience and referral-chain sampling approaches, two commonly used non-probabilistic sampling methods (Pechansky et al., 2004; Bolfarine and de Oliveira Bussab, 2005; Mendenhall et al., 2006). Convenience sampling is a sampling method in which units are selected based on easiness of access or availability. This method of sampling is generally less time-consuming and of lower cost. Referral-chain sampling is a technique that is often used for intentional selection of expert informants. In this study, we were only interested in respondents who have produced the species for at least 5 years, and/or who consumed/exploit its fruits. In this method, the first contact with the community may be through a well-known expert who then indicates another expert, and so on, until the sample size is reached (Albuquerque et al., 2010). The entry points in our case were the community chiefs or leaders. Combining these two sampling strategies was relevant to our study since we did not have any prior information on the species owners, and interestingly the sampling threshold imposed by the saturation point (defined as the point at which there are no new names of key informants being mentioned by the last interviewee; N’Danikou et al., 2015) was observed to represent enough the population studied.

Data were collected through formal semi-structured interviews with respondents using a questionnaire. During the exercise, information describing the respondents’ socio-demographic and cultural characteristics (sex, age, and sociolinguistic group affiliation), farm characteristics, cultivation practices of Sisrè berry plant, awareness of existence of varieties or morphotypes of the species, farmers and consumers’ preferences for breeding traits as well as their production constraints.

Analysis of data involved the use of descriptive statistics such as frequency, percentages, standard deviation, and mean, to describe the socio-economic characteristics of respondents. The mean values and the standard deviation were calculated on farmers and consumers preferences scores. The χ² Fisher exact test was performed to assess the dependence relationships between (i) the awareness of a specific morphotype and the sociolinguistic group affiliation and (ii) the respondent’s involvement in Sisrè berry cultivation and their gender /education level /main activities. The ANOVA with least significant different (LSD) post hoc test was performed using “agricolae” package (de Mendiburu and de Mendiburu, 2019) to assess the statistical grouping in informants’ age, household size, total farm size (acre), and farmers experience in Sisrè berry plant management (years). The factorial analysis of mixed data was performed to describe Sisrè berry plant farming systems and establish the typology of the farmers while the Factorial Analysis of Correspondence (FAC) was performed using “FactoMineR” package (Lê et al., 2008) to test the relationship between the listed preferences and sociolinguistic groups, regions, and gender. The mean value and the standard deviation were calculated for scored constraints across regions and sociolinguistic groups. The Kendall’s Coefficient of Concordance (KCC) was computed using “irr” package (Gamer et al., 2012) to test the level of agreement among the listed constraints. The value of KCC is positive and ranges from 0 to 1. All the analyses were carried in the R environment (version 4.1.2; R Core Team, 2021).

A total of 100 respondents from four States including Ondo, Oyo, Anambra, and Enugu were interviewed in this study. Across these four States, a greater proportion of male farmers was involved in the Sisrè berry plant cultivation compared with women (Table 1). The greatest proportion of women involvement (36% of respondents) was observed in Enugu state. Respondents from Anambra and Enugu belong to the “Igbo” sociolinguistic group, while those from Ondo and Oyo are “Yoruba.” Informants from Oyo and Enugu were the most literate (68 and 16% for tertiary and secondary levels, respectively, in Oyo region; and 36 and 32% for tertiary and secondary levels, respectively in Enugu). The highest proportion of respondents with no formal education was observed in Ondo. Agriculture (p < 0.0001, χ² = 28.04, df = 3) and trade (p < 0.0001, χ² = 42.41, df = 3) were the main activities of the respondents with high significant difference across the states. The respondents were on average 47.8 ± 8.5 years old, with respondents in Oyo being the youngest (44.5 ± 8.1 years old), while those from Enugu were the oldest (50.2 ± 9.9 years old). Informants from Anambra and Enugu states had a greater experience (26.1 ± 6.5) in the Sisrè berry plant management (Table 1).

The Sisrè berry plant was mainly found in home gardens although also existing in cultivated farms, fallows or naturally growing in forests. Most of Sisrè berry plant owners also possessed plantations of other perennial species including for instance cashew (Anacardium occidentale L.), cocoa (Theobroma cacao L.), coconut (Cocos nuciferae L.), palm trees (Elaeis guineensis L.), among others. None of the farmers applied all required management practices in the production of the species (land clearing before crop establishment, pegging and lining, holing, base manure application, crop irrigation, weeding, pruning, fertilizers application, and pest and diseases control).

The factorial analysis of mixed data grouped the farmers into three (03) categories based on the various farming practices they applied (Figure 3). The distribution of variables characterizing these various categories is presented in Figure 4. The cluster 1 contained 93.6% of the interviewed farmers. Famers in this cluster owned on average two stands of Sisrè berry plant and had 18.50 ± 8.8 years of experience in the species management. None of them attended a training in biodiversity conservation and 97.7% of them possessed individual trees of Sisrè. Trees were mainly inherited (59.1%) and dispersed in home gardens or cultivated farms. Seeds were obtained from the ripe fruits and the regeneration mode they applied included: seeds sowing (97.9%), seedlings transplanting (82.9%), and cuttings (23.4%). A small proportion of these farmers (21.3%) applied base manure before transplanting the seedlings and irrigated the transplanted seedlings (44.7%). Farmers mostly applied the weeding (65.9%) while the chemical pesticides spray was only applied by 2.1% of them. This cluster included 22.7% of farmers from Ondo, 31.8% of farmers from Oyo (54.5% of Yoruba), 18.2% of farmers from Anambra, and 38% of farmers from Enugu (45.5% of Igbo).

The cluster 2 contained 2.1% of farmers (one respondent) belonging to the Yoruba sociolinguistic group (Oyo). This farmer was the only one to possess a plantation of Sisrè berry plant with up to 2,700 stands. This farmer established the Sisrè berry plant orchard by transplanting the seedlings produced by sowing matured fruits. Required management practices such as basal dose of manure, irrigation, weeding, and phytosanitary spraying were done.

The cluster 3 encompassed 4.2% of farmers from Oyo (Yoruba) with a moderate number of Sisrè berry trees (up to 115 stands). They have attended some training in biodiversity conservation and established the plantations themselves. The land clearing, pegging, lining, holing, and base manure application were systematically carried out before transplanting the seedlings. Crop management included weeding and irrigation. There were no phytosanitary products spraying by farmers in this cluster (Figure 4).

None of the surveyed farmers in Southern Nigeria differentiated any variety of S. dulcificum. However, they recognized two morphotypes based on the color of the fruit exocarp: a red-skinned morphotype (Figure 5A) and a yellow-skinned morphotype (Figure 5B). There was a strong association between awareness of different morphotypes and sociolinguistic group affiliation (p < 0.0001, χ² = 96.0, df = 1) and the agroecological region (p < 0.0001, χ² = 96.0, df = 3) with the yellow morphotype exclusively reported by the Igbo in the Southeast region and the red-skinned morphotype only known by the Yoruba, in the Southwest. In both regions, the species is becoming less and less available despite having two fruiting seasons yearly. Seventy-four percent (74%) of the respondents confirmed the scarcity of the species in Southeast as compared to 62% in Southwest (p = 0.30, χ² = 1.0, df = 1). Several reasons justified this state of availability including bush fire during the dry season, destruction of the habitat of the species for settlement, road construction or agriculture expansion, the poor rate of the plant growth, and the fact that the species is not yet entirely domesticated and that cultivation in agricultural production systems remained low (Figure 6).

In Southeast Nigeria, inhabited by Igbo, the top five constraints identified by the Sisrè berry plant growers included insect and birds’ damages, poor knowledge of cultivation technics, drought, and low seed germination (Table 2). While farmers in Anambra and Enugu had three (03) out five of these constraints in common, there was only a moderate agreement in their ranking of the constraints (Kendall’s W = 0.6, p < 0.001). In the Southwest region inhabited by the Yoruba, the five most important constraints included insect and birds’ damages, drought, bushfire, and low seed germination. There was a higher concordance in the ranking of the constraints between farmers in Ondo and Oyo (Kendall’s W = 0.8, p < 0.001) compared with their peers in Anambra and Enugu (Kendall’s W = 0.6, p < 0.001). Overall, there was a good agreement for the constraints list and ranking between the surveyed Igbo and the Yoruba farmers (Kendall’s W = 0.6, p < 0.001; Table 3).

The harvested products were mainly for household use (63.2%), followed by sales (21.3%) and gifts (15.5%). The fruits sales mainly occurred on local market (94.3% of respondent selling the fruits), while only a low proportion of harvested fruits was processed into powder or juice and sold in capital city Lagos or exported to the United States (Table 4). There was not specific selling unit for sales in the local market while the kilogram is used for the international market.

Producers and consumers desired a range of traits for an improved variety of S. dulcificum. A total of 11 and six traits were mentioned by farmers and consumers, respectively. For farmers, the top five most preferred breeding traits for S. dulcificum included by descending importance, a high plant growth rate, an expanded fruit shelf -life, a high fruit yield, a high pest and drought resistance or tolerance, and a high fruiting frequency. The most important traits for consumers in descending importance included a longer sweetening action of the fruit, a higher fruit glossiness, higher metabolites content, a thicker pulp size, and a larger fruit size (Table 5). All these desired attribute were in comparison with their local landraces’ performance.

The factorial analysis of correspondence revealed a variation of farmers’ preferred traits across sociolinguistic groups and gender. For Igbo, the priority breeding traits included fruits size, early fruiting, seed viability, and germinability, while Yoruba rather mainly targeted fruits yield, fruits shelf life, fruits glossiness, plant growth rate, tree life span, and the fruiting frequency. The gender exerted a lesser influence on farmers’ preferences compared with sociolinguistic groups, with men prioritizing fruits glossiness, fruits shelf life, early fruiting, and seed viability while women did not show any specific preference (Figure 7). There was also a larger variation of trait preferences within Igbo compared with Yoruba farmers (Figure 8). Noticeably, the Anambra respondents had a higher interest for traits such as early fruiting, fruit glossiness, and seed viability, while their counterpart from the Enugu were more interested in fruits shelf life and seed germination (Figure 8). Consequently, a higher congruity was found in farmers’ preferences in Southwest also known as Yoruba land (Kendall’s W = 0.9) compared with a moderate congruity in the identified preferences in the Southeast known as Igbo land (Kendall’s W = 0.5; Table 6).

Consumer’s preferences varied across ecological zones (Figure 9A) on one hand, and across sociolinguistic groups and gender (Figure 9B) on the other hand. Consumers from Southeast region, the Igbo, valued the glossiness of fruits (fruits aspect) and its metabolites content while those in the Southwest, the Yoruba, preferred a long sweetening action of the fruit, the taste, the fruits size, and the fruits shelf life. There was no variation in consumer’s preferences within the Southeast region (Anambra and Enugu states). Conversely, in the Southwest region, while the informants from Ondo state preferred the fruit shelf life and the fruits size, their peers in Oyo state preferred the fruit taste and its potency (long lasting of the sweetening activity). There was also a significant concordance in consumers preferences in Southeastern Nigeria (Kendall’s W = 0.7), while a weak congruity was found in the identified preferences in Southwestern Nigeria (Kendall’s W = 0.4; Table 6). Consumers’ preferences were also gendered with women prioritizing fruit taste, fruit shelf life, and metabolites content whereas men were rather interested in traits including the fruit glossiness, potency, and the fruit size (Figure 9B).

This study, conducted in Nigeria, expanded our knowledge of the current habitats of the Sisrè berry plant in West Africa. Previous studies reported the species to be predominantly present in-home gardens, cultivated farms, fallows, and gallery forests (Fandohan et al., 2017; Tchokponhoué et al., 2021). In this study, the species was also found in rainforests, which reflected that the suspected habitat fragmentation in the species may not be evolving at a similar rate across the distribution range in West Africa. Home gardens are overall becoming a back-up habitat for the species and serve as a reservoir for biodiversity conservation (Catalán et al., 2007; Galluzzi et al., 2010; Fandohan et al., 2017). The farmers typology pointed out to three groups with most of farmers (cluster 1 members) applying only limited practices, a result that corroborated observations in the republic of Benin where management intensity index was low (Management intensity Index < 15 vs. >30 in the evergreen region of Ghana; Tchokponhoué et al., 2021). The overall poor management of the species (from cluster 1 members) could be explained by the low economic value ascribed to the species by respondents in this cluster. Such a perception is not likely to trigger any active production initiative, and this was illustrated by the high tendency to inherit the species, a result that is also congruent with several previous studies on the dominant mode of acquisition of the Sisrè berry plant (Fandohan et al., 2017; Tchokponhoué et al., 2021). Only, a low proportion of respondents (cluster 2 and cluster 3) considered the species as economically important. Noticeably, those respondents mostly had a high level of education, which explained their motivation for large-scale cultivation of the species (up to 2,700 stands). These farmers have a better access to the information about the crop, in particular the trading opportunities in the crop. Indeed, these farmers are anticipating on the potential value of the species and are targeting either the international niche market (cluster 3) or have local market to satisfy (Cluster 1). In Ghana already, it was noticed that any single farmer holding a plantation of the Sisrè berry plants has a contract with a processing company that not only guarantees the market, but also provides funding for the plantation maintenance (Tchokponhoué et al., 2020). This calls for the necessity for implementing a subsidy system for the production of the Sisrè berry plantations, should West African governments promote this tremendous resource.

The two cultivars of Sisrè berry reported in this study were previously signaled by various authors (Njoku et al., 2015; Tchokponhoué et al., 2020). While the red-skinned morphotype is widely distributed across Africa (Inglett and May, 1968; Bartoshuk et al., 1974; Huang et al., 2012; Njoku et al., 2015; Iloh et al., 2017) and well known worldwide (Inglett and May, 1968; Bartoshuk et al., 1974; Huang et al., 2012; Njoku et al., 2015; Iloh et al., 2017), the yellow-skinned morphotype seemed to be confined to Southeast Nigeria in West Africa (Inglett and May, 1968; Bartoshuk et al., 1974; Huang et al., 2012; Njoku et al., 2015; Iloh et al., 2017). Indeed only the red morphotype was reported by famers investigated in Benin, Ghana, and Togo in the framework of the ongoing breeding initiative on the species (Tchokponhoué et al., 2020). The fact that the two morphotypes reported by farmers are each specific to a region could be explained either by a low or absence of exchange of planting material of the species between Igbo and Yoruba or the hypothesis of habitat preference by each morphotype. Angeles et al. (2017) already reported an emergence of the yellow-fruited morphotype in a lot of 23,000 Sisrè berry plants cultivated in Calauan, Laguna, Philippines though where these plants batches were obtained from was unclear. Nevertheless, since the Sisrè berry plant is known to originate from West Africa, we can speculate that Southeast Nigeria would the center of origin of the yellow morphotype. In the context of the breeding strategies in the species, this yellow morphotype represents a potential source of favorable alleles that can be tapped to improve some key traits of interest for the farmers. Indeed, the fruit of the yellow morphotype is nearly two times bigger (data not shown) than the red morphotype and can for instance be relevant in the development of big fruit-sized cultivars. In addition, the yellow morphotype exhibited some primitive characters such as prominence of pubescence in several plant parts (fruits, leaves), seed coat thickness (data not shown). Consequently, further in-depth comparative phenotypic analysis and molecular phylogenetic analysis are then needed to clarify the relationship between these supposed morphotypes.

Insect and bird damages and difficulties in the crop management were the major constraints faced by the surveyed farmers in Southern Nigeria. In fact, the ripe miracle fruit is a small berry finely pubescent with bright red/yellow skin. The brightness combined with the fruit’s sweetness attracts herbivorous insects and birds which feed on the fruits (Stevens and De Bont, 1980). These insects and birds-induced damages were previously reported in cherries (Lindell et al., 2012), peach, apple, pear, grape, and loquat cultivation (Hao et al., 2011; Lindell et al., 2012) and are likely to induce important economic impact on producers (Anderson et al., 2013; Angeles et al., 2017). As a physical protection technique, pre-harvest bagging of fruits was developed and have been chiefly used in several species including Vitis vinifera L. (grapevines; Karajeh, 2018), Psidium guajava L. (guava; Srivastava et al., 2023), and Mangifera indica L. (mango; Nadeem et al., 2022). The pre-harvest bagging optimizes fruit quality by reducing physical damages and damages by pathogens, and increases market value of the fruits (Sharma et al., 2013).

Quick loss of seed viability and poor germination seemed to be more common in Southeast, while slow plant growth, drought, and bushfires were prominent in the Southwest region. It has been established that S. dulcificum produces recalcitrant seeds with seeds losing viability in few days after harvesting (Tchokponhoué et al., 2019). This explains farmers’ observation of quick loss of viability and poor germination. A slow growth of plants was reported by over 70 % of respondents. This observation was experimentally established by Tchokponhoué et al. (2018). These observations suggested farmers’ good knowledge of the species’ biology. Other studies also reported Sisrè berry plant as a slow-growing species with two growth phases. The first growth phase corresponding to the first 4 years where the plant grows very slowly reaching about 50 cm tall, and a second phase, starting from 4-year-old onwards where the plant grows faster. Synsepalum dulcificum’s growth is slow compared to other economically important species. For instance, at the same age, Vitellaria paradoxa (a sister species to S. dulcificum in the Sapotaceae family) can be four times taller (Allaye Kelly et al., 2004). The higher frequency of seed-biology related challenges reported in the Southeast indirectly suggested that the yellow-skinned morphotype could have a more sensitive seed physiology compared with the red-skinned, since producers of the later morphotype did not face too much such a challenge. This poor seed physiology marked by a short seed lifespan of the yellow-skinned morphotype might explain its low popularity. Drought and bushfires are important constraints in the Southwest of Nigeria. Indeed, in this region, the species is mainly found in farms, fallows, or forests, which are open-habitats, hence low protection and management by farmers as compared with plants found in home gardens in the Southeast. Bushfires were commonly reported to threat for several economically important trees such as Vitellaria paradoxa C. F. Gaertn., Tamarindus indica L., and Sclerocarya birrea (A. Rich.) Hochst (Gaisberger et al., 2017). The dry season in southern Nigeria lasts for 6 months (November to March) yearly (Ojo, 1977), a period over which local communities used to burn the bush and some parts of forests which are also habitats of Sisrè berry plant.

The harvested products were mainly for household use or sold on local markets by about 94% of respondents. However, the other 6% of the farmers have access to the international market, with an important export of fresh fruits to United States. This more advanced farmers also explore the potential for processing fruits into powder or juice locally. This indicates that there is a potential to up-scale production as market opportunities exist. An illustration is the European Union market that has recently admitted the dry powder of the Sisrè [EFSA Panel on Nutrition, Novel Foods and Food Allergens (NDA) et al., 2021]. Besides fruit uses, the dry seeds have high commercial value as the dry seeds attract approximately two times higher price compared with the fresh fruits. The dry seeds are used in the production of a cosmetic oil used in women hair breakage treatment (Del Campo et al., 2017).

An analysis of end-users’ preferences helps breeders detect hidden traits of interest, and both famers and consumers formulated their trait preferences (Bolfarine and de Oliveira Bussab, 2005). For instance, the slow growth was on the highlighted challenges, indicating the desire of farmers to have a fast-growing cultivar that can bear fruits earlier. Although farmers in Nigeria listed a lower number of preferred-traits (11) compared with their counterparts in Benin and Ghana (who highlighted 19 traits), it is worth pointing out that seven out of the 11 traits were already reported by farmers in Benin and Ghana (Tchokponhoué et al., 2021). This suggest that Sisrè berry plant farmers in West Africa share preferences for an improved variety of Sisrè berry plant. More importantly, four out of the top five desired traits in this study were previously highlighted by the species producers from Ghana and Republic of Benin (Tchokponhoué et al., 2021), thus strengthening the necessity to promote region-wide cultivars development initiatives.

Most of the farmer’s preferences converged to high fruit yielding while consumers preferred visual and nutritional traits. Farmers in miracle fruits production are seeking to maximize the profit through increased fruits yield while consumers would like to enjoy quality fruits with high glycoproteins and metabolites contents. End-users’ preferences are most important to increase adoption of new varieties. To illustrate, most farmers in Africa and Latin America keep growing some specific Andean type of rice varieties with lower yield to satisfy particular traits required by consumers, even though some high-yield improved varieties (Mesoamerican types) exist (Beebe, 2012). Our findings revealed a variation of farmer’s preferences across regions, sociolinguistic groups, and gender. However, from a West Africa regional perspective, combining previous findings (Tchokponhoué et al., 2020) with the current ones will lead to the definition of key breeding traits for both farmers and consumers, and the establishment of product profiles by breeders.