Assessment of the roles and farmer-preferred traits of sweetpotato in a crop-livestock farming system in Rwanda: implications for breeding dual-purpose varieties

In Rwanda, sweetpotato (Ipomoea batatas [L.] Lam) is a highly valued crop in a crop-livestock mixed farming system. The objective of this study was to assess the role of sweetpotato in the crop-livestock farming system, to identify farmer-preferred traits, and to establish farmer-led priorities in breeding dual-purpose varieties (DPVs) in Rwanda. A participatory rural appraisal (PRA) study was conducted in three selected districts of Rwanda, namely, Bugesera, Huye, and Nyagatare. Data on the uses of sweetpotato and farmers’ trait preferences in sweetpotato varieties were collected and analysed. In Huye District, a high percent (56.7%) of respondents consumed sweetpotato every day, followed by Nyagatare with 53.3% consuming it at least twice a week. Most farmers (52.2%) used sweetpotato vines for livestock feed, depending on their availability. All respondents wanted to grow new sweetpotato varieties with improved root production combined with high aboveground biomass. About 87.7, 66.6, 56.6, and 51.1% of the respondents indicated that root-related traits of the crop such as high dry matter content (DMC), red skin colour, marketable root size, and yellow flesh colour were additional preferred traits, respectively. Therefore, farmers-preferred DPVs with improved root and green fodder yields could be developed to enhance the sustainable production and adoption of sweetpotato in a mixed farming system in Rwanda.


Introduction
Sweetpotato [Ipomoea batatas (L.) Lam] is an important crop in many regions of the world. Asia and Africa are the predominant sweetpotato producing continents, contributing to 76.4 and 19.2% of the world annual production, respectively. In Rwanda, sweetpotato is the second important crop after bananas, with the total annual production of 10,82,364 tonnes (FAOSTAT 2018). Sweetpotato has multiple uses including food, animal feed, and recent breeding efforts focused on dual-purpose varieties (DPVs) (Shumbusha et al. 2019; Rukundo et al. 2020). It provides more edible energy per hectare than other food security crops such as maize, wheat, and rice (Mukhopadhyay et al. 2011). The orange-fleshed sweetpotato varieties (OFSPs) are rich in beta-carotene content, the precursor of vitamin A, an essential nutrient (Ginting 2013;Grüneberg et al. 2015). As the commercialization of storage roots increases, fresh roots with a good shape and complying with other market requirements are produced for export. In developing countries, sweetpotato is mainly cultivated for its storage roots for human consumption and its aboveground biomass for livestock feed, making it an ideal dual-purpose crop (León-Velarde 2000; Ralevic et al. 2010; Valbuena et al. 2012).
The dual-purpose nature of sweetpotato is highly valued in crop-livestock mixed farming systems in sub-Saharan Africa, including that practiced in Rwanda where the availability of agricultural lands is severely limited (Kamanzi and Mapiye 2012; Shumbusha et al. 2017; Shumbusha et al. 2019). Rwanda has a high population density estimated at 430 inhabitants km −2 (FAOSTAT 2018). Sweetpotato is grown across varied agroecological zones in smallholder farming systems. The most suitable agroecologies for cultivating the crop are the low and mid-altitudes including the semiarid areas of Bugesera and Nyagatare, which are situated in the Eastern Province of Rwanda. The crop is produced under rainfed conditions with a bimodal rainfall pattern across the country. Consequently, most sweetpotato smallholder farmers plant and harvest around at the same time as they must match the production cycles. During planting, farmers prepare the soil manually with a hoe and planting is done either on ridges or on flat, with 30 cm within rows and 80 cm between rows. The mean on-farm yield of sweetpotato in Rwanda is 6.5 t ha −1 which is below the attainable mean yield of 14.2 t ha −1 (NISR 2017). The yield gap is attributed to biotic and abiotic stresses and socioeconomic constraints that vary across agroecologies in the country. Socioeconomic factors limiting sweetpotato production in Rwanda include: limited agricultural land; shortage of clean and adequate planting material; and poor agronomic management practices (Niyireba et al. 2013;Makokha et al. 2020).
Farming in Rwanda is characterized by fragmented small plots of land measuring less than one hectare per household (MINAGRI 2010). Each household involves in a number of interdependent farming enterprises, predominantly in crop and livestock production. The farming systems in the country are highly heterogeneous due to variable household resources (Dixon et al. 2001). Therefore, any strategy to develop these farming systems should consider the prevailing farming practices, farmer's constraints, and the overall socioeconomic aspects (Bucagu et al. 2013). A research approach to document farmers' circumstances and constraints should involve farmers' participatory methods across the farming systems (Almekinders and Elings 2001).
Participatory research techniques have been used successfully to identify farmers' perceived production constraints, preferred crop varieties, and key traits for deployment of production packages and the development of suitable crop varieties (Ndolo et al. 2001;Ndirigwe et al. 2005;Tefera et al. 2013). Varieties like Kwezikumwe, Mugande, SPK004, and Wagabolige have shown their potential to produce high storage root and aboveground biomass yields, but varieties with strong dual-purpose ability were not yet officially released (Niyireba et al. 2013). Among participatory research techniques, the participatory rural appraisal (PRA) approach has been successfully used in identifying farmers' constraints, preferred traits, and needs (Chiona 2010;Muhinyuza et al. 2012;Kivuva et al. 2014). Therefore, the objectives of this study were to assess the role of sweetpotato in the crop-livestock farming system, to identify farmer-preferred traits, and to clarify the objectives in breeding DPVs in Rwanda.

Study sites
The study was conducted across three selected districts of Rwanda, namely, Huye, Nyagatare, and Bugesera ( Figure 1). These districts are known for their sweetpotato production. Huye District (02°29′S, 29°46′E) is situated in the Southern Province with an altitude of 1,700 m above sea level [masl]. Huye has an average temperature of 18.7°C with a total annual rainfall of 1,200 mm. Nyagatare and Bugesera Districts are situated in the Eastern Province and known for their mixed crop-livestock farming systems. Bugesera (02°17′ S; 30°16′ E) is located in the lowlands (<1,400 masl) with a total annual rainfall of 700-900 mm and a mean temperature of 20.8°C (Murayi et al. 1987). Bugesera is a hotspot area for sweetpotato virus diseases (SPVD). Nyagatare District (1°22′ 51.6″ S; 30°17′ 07″ E) is located in the East Savanna agroecological zone with an altitude of 1,400 masl (Nabahungu and Visser 2011). This district ranks first in livestock production in the country.

Sampling procedure and data collection
A purposive sampling (Frankel and Devers 2000) was used in order to increase the likelihood of including relevant sites and samples in the study. The following six administrative sectors were selected in the three districts: Mwogo, Ntarama, and Rweru (Bugesera District), Huye (Huye District), and Katabagema and Tabagwe (Nyagatare District). In each sector, two administrative cells were sampled, resulting in a total of 12 cells. Two villages per cell were sampled, providing a total of 24 villages for the study. In each village, three to eight farmers were sampled. This provided a total of 90 farmers sampled using semi-structured interviews. Focus group discussions (FGDs) were held involving nine focus groups comprising farmers, local leaders, and key informants. Each focus group was composed of six to ten representative farmers who were sampled based on their experience in sweetpotato production. A total of 78 farmers participated in the FGDs across the three districts. A multidisciplinary research team was constituted for the study. The team composed of a sweetpotato breeder, two research technicians, a socioeconomist, an animal nutritionist, and local key informants.
Data were collected through semi-structured interviews, focus group discussion, and a transect walk. In addition, secondary data were collected from previous reports. Semi-structured questionnaires were used to collect data from individual farmers. Data collected through semi-structured interviews included the importance and uses of sweetpotato, constraints to sweetpotato production, farmers' interest in DPVs, and consumer preferences. FGDs were held to gather information such as importance of the crop in food security and ranking of the currently grown sweetpotato varieties and their characteristics. FGDs also included pair-wise and matrix ranking of the crops grown. Gender balance was taken into consideration by involving both male and female farmers during the semi-structured interviews and focus groups. The gender groups were essential in order to collect data on role of males and females in sweetpotato production and postharvest activities.

Data analysis
Data collected were analysed using Statistical Package for Social Sciences (SPSS) for windows Release version 21 (SPSS 2012). The analyses involved descriptive statistics and cross-tabulations in order to calculate percentage of respondents for each question or focus group. Chi-square tests were computed in order to determine associations between collected parameters and the study districts, and therefore, to make statistical inferences.

Sources of incomes of households and forms of land ownership
Crop production activities were the main source of income for 60 and 73.3% of the respondents in the Bugesera and Huye Districts, respectively. A mixed croplivestock system was the main source of income in the Nyagatare District, with 76.7% farmers pursuing their livelihood in this sector ( Table 1). Sources of income varied significantly across districts (X 2 = 16.121, P = 0.001). About 70, 93.3, and 100% of the respondents owned farms in Huye, Bugesera, and Nyagatare Districts, respectively. However, in the Huye District, 26.7% of respondents rented farms ( Table 1). Land tenure varied significantly (X 2 = 14.096; P = 0.007) between the study districts. Approximately 45% of the interviewees had farms of less than 0.5 ha. In the Nyagatare District, 80% of the respondents had land holdings of 2 ha or more.

Uses of sweetpotato in the farming systems
About 80, 86.7, and 96.7% of the respondents used sweetpotato vines as a green fodder for livestock feed in the Huye, Bugesera, and Nyagatare Districts, respectively ( Table 2). A limited number of respondents (10.2%) mentioned the use of vines as planting material, though a relatively high percentage (20%) was reported in the Huye District. Vine-based fodder was commonly used in Nyagatare District, reported by nearly all the respondents (96.7%). There were nonsignificant differences for the use of sweetpotato vines across the three surveyed districts.

Farmers' strategies to increase sweetpotato productivity
Most respondents perceived that crop improvement would be the best strategy to increase sweetpotato root productivity. This was stated by more respondents (93.3%) in the Huye District than respondents in the Bugesera and Nyagatare Districts (76.7%) ( Table 3). Farmers' perceptions regarding strategies to increase root production were relatively similar across districts.
Among respondent farmers, about 86.7% in Nyagatare and 90% in both Bugesera and Huye Districts believed that improved sweetpotato varieties could bring about high levels of fodder production, whereas only 10% of respondents in Bugesera and Nyagatare Districts proposed better crop management options ( Table 3). Perceived strategies to increase fodder production did not vary across districts.

Farmers-preferred sweetpotato varieties
Most respondent farmers expressed a desire for productive dual-purpose sweetpotato varieties in preference to specifically storage root or vine types. In the Huye District, most farmers (96.7%) would prefer DPVs, compared to Nyagatare and Bugesera Districts with 80 and 83.3% of farmers, respectively ( Table 4). The proportions of respondents on the level of sweetpotato preferences were not significantly different across districts.  Table 5). Other constraints reported were diseases, shortage of planting material, and drought. Nonsignificant differences were detected between districts and sectors for the reported constraints.    Preferences of sweetpotato varieties Respondents (%) Root production 2 (6.7) 2 (6.7) The top two main constraints to sweetpotato production identified through focused group discussion were SPVD and a lack of planting material. SPVD was described as a major constraint ( Table 5) in all three surveyed districts. A lack of DPVs and poor soil fertility were the overriding constraints specific to Huye District, whereas weevils and a lack of postharvest facilities were reported as the main constraints in Nyagatare District.

Major characteristics of farmerspreferred sweetpotato varieties for storage root production
About 50 and 76.6% of farmers preferred marketable storage roots, in the Bugesera and Huye Districts, respectively, whereas in the Nyagatare District the farmers wanted medium-sized roots (    Poor soil fertility -6 (1) -1 9 -: Not ranked among the preference crops grown in the district and data did not allow to conduct significant tests.
The roles and farmer-preferred traits of dual-purpose sweetpotato varieties in Rwanda  839 This study investigated the varied role of sweetpotato in Rwanda as a food and fodder crop. It also determined key farmer-preferred traits of the crop as a guide to subsequent breeding of dual-purpose varieties. The current findings confirmed the value of sweetpotato as a dualpurpose crop in Rwanda, given the limited agricultural land available for crop and livestock production in the country.

Household information
Crop production was the main source of income in the Bugesera and Huye Districts, whereas a mixed crop-livestock system characterized the farming system in the Nyagatare District (Table 1). This showed the overall importance of crop production in the country farming systems. The mixed crop-livestock farming in the Nyagatare District reflected the high level of livestock production in this district. Similar results on the important role of crop production in Rwanda were also reported by Muhinyuza et al. (2012). Land tenure varied significantly ( Table 1). Although access to land is a challenge in the country, plot size, land availability, and accessibility were different across the surveyed districts.

Role of sweetpotato in the present farming systems
The current findings indicated that sweetpotato storage roots were commonly used for food ( Table 2)

Strategies to increase sweetpotato productivity
Farmers' perceptions regarding strategies to increase root production were similar across districts ( Table 3). Respondents perceived that cultivar improvement would be the best way to increase sweetpotato root productivity, concurring with the previous results of Rukundo et al. (2015). This reflects the need to develop and disseminate new varieties with improved root yields. Similarly, most farmers expressed interest towards new varieties with greater fodder production potential ( Table 3).

Preferences of dual-purpose sweetpotato varieties
Most respondents in all districts expressed their need for DPVs, especially in Huye District (

Constraints to sweetpotato production
There were several constraints to sweetpotato production such as: pests, diseases, shortage of clean planting material, lack of DPVs, and drought (

Major characteristics of the preferred storage root
In Bugesera and Huye Districts, 50 and 76.6% of farmers preferred marketable root size of storage roots, respectively ( Table 6). This reflects the commercial value of the crop, which is steadily increasing across districts. Relatively few farmers preferred marketable roots in Nyagatare District because of the commercial value of other crops such as maize and rice. Most farmers preferred red skin colour. Similarly, Ndirigwe et al. (2005) reported that red skin colour was consistently liked over time in Rwanda. In the Bugesera and Nyagatare Districts, white-fleshed sweetpotato varieties were preferred over cream, yellow, and orange flesh colour ( Table 6). The study of Hagenimana et al. (1998)  Breeding of sweetpotato in the study areas should take into consideration these farmers' preferred traits.

Conclusions
The study revealed the importance of sweetpotato as a food security crop with dual-purpose potential in Rwanda. All respondents expressed interest in growing sweetpotato varieties with improved root production combined with high aboveground biomass. About 87.7, 66.6, 56.6, and 51.1% of the respondents indicated that root-related traits of the crop such as high DMC, red skin colour, marketable root size, and yellow flesh colour were additional preferred traits, respectively. Therefore, farmers-preferred dual-purpose sweetpotato varieties with improved root and green fodder yields should be developed to enhance sustainable production and the adoption of sweetpotato in the mixed farming systems in Rwanda.