A challenge for African countries is how to integrate new sources of knowledge on plant genetics with knowledge from farmer practice to help improve food security. This paper considers the knowledge content of farmer seed systems in the light of a distinction drawn in artificial intelligence research between supervised and unsupervised learning. Supervised learning applied to seed systems performance has a poor record in Africa. The paper discusses an alternative-unsupervised learning supported by functional genomic analysis. Recent work in West Africa on sorghum, African rice and white yam is described. Requirements for laboratory-based analytical support are outlined. A science-backed 'farmer first' approach-while feasible-will require a shift in policy and funding by major investors.

The maintenance and utilization of crop genetic diversity is important to ensure food security. The relative importance of yam and cowpea varieties and the influence of the socio-cultural and local economy context on the diversity maintained were analysed in Benin. Whereas the diversity is large, some varieties were rare, other ones on the way of being abandoned or already lost. Socio-cultural as well as economic and agronomic characteristics explained why some of them were still maintained. For example, the early-maturing yam variety Laboko was planted by most farmers to have tubers available in time for religious purposes, and some specific cowpea varieties played a role in the funeral of the parents in law. Farmers' preferences were translated into criteria they use to appreciate varieties. The diversity of the varieties sold on the market and their availability over time reflect farmers' strategies and conservation practices. The large price differences between varieties confirm the variation in quality as perceived by consumers. The most widely grown yam variety, Florido, is available on the market throughout the year but has a very low price. Market price differences among varieties are much smaller for cowpea than for yam. The processes of loss and displacement of some local varieties are described and the need for conservation is addressed. Different factors that may influence the level of varietal diversity in these crops, like the need to synchronize harvesting with high market prices, were analysed in depth. As opposed to mono-disciplinary approaches of scientists to farmers' problems and constraints, farmers show an inter-or trans-disciplinary behaviour and express their preferences through multi-criteria processes.

Spodoptera frugiperda has caused significant losses of farmer income in sub-Saharan countries since 2016. This study assessed farmers’ knowledge of S. frugiperda, their perceptions and management practices in Benin. Data were collected through a national survey of 1237 maize farmers. Ninety-one point eight percent of farmers recognized S. frugiperda damage, 78.9% of them were able to identify its larvae, and 93.9% of the maize fields were infested. According to farmers, the perceived yield losses amounted to 797.2 kg/ha of maize, representing 49% of the average maize yield commonly obtained by farmers. Chi-square tests revealed that the severity of the pest attacks was significantly associated with cropping practices and types of grown maize varieties. About 16% of farmers identified francolin (Francolinus bicalcaratus), village weaver (Ploceus cucullatus), and common wasp (Vespula vulgaris) as natural enemies and 5% of them identified yellow nutsedge, chan, shea tree, neem, tamarind, and soybean as repellent plants of S. frugiperda. Most farmers (91.4%) used synthetic pesticides and 1.9% of them used botanical pesticides, which they found more effective than synthetic pesticides. Significant relationships exist between farmers’ management practices, their knowledge, organization membership, and contact with research and extension services. More research is required to further understand the effectiveness of botanical pesticides made by farmers against S. frugiperda and to refine them for scaling-up.

PurposeThis document analyses farmers' preferences and willingness to pay (CAP) for microcredit, in order to facilitate their access in rural areas.Design/methodology/approachData are based on a discrete choice experiment with 400 randomly selected farmers from 20 villages of the 7 Benin agricultural development hubs (ADHs). The preference choice modelling was performed using mixed logit (MXL) and latent class logit (LCL) models. Farmers' willingness to pay for each preferred attribute was estimated. The endogenous attribute attendance (EAA) model was also used to capture attribute non-attendance (ANA) phenomenon.FindingsThe results indicate that, on average, farmers prefer individual loans, low interest rates, in kind + cash loans, cash loans, disbursement before planting and loans with at least 10-month duration. These preferences vary according to farmers' classes. Farmers are willing to pay higher or lower interest rates depending on attribute importance. The estimate of the EAA model indicates that, when taking the ANA phenomenon into consideration, people will show stronger attitudes regarding WTP for important factors.Research limitations/implicationsBased on these results from Benin, microfinance institutions (MFIs) in developing countries can, based on the interest rates currently charged, attract more farmers as customers, reviewing the combination of the levels of the attributes associated with the nature of the loan, the type of loan (individual or collective), the disbursement period of funds, the waiting period of the loan and the loan duration. However, the study only considered production credit, ignoring equipment or investment credit.Practical implicationsThe document provides information on the key factors that can facilitate producers' access to MFI products and services.Social implicationsFacilitating small farmers' access to financial service will contribute to poverty reduction.Originality/valueThis research contributes to the knowledge of the attributes and attribute levels favoured by farmers when choosing financial products and the amounts they agree to pay for these attributes. The implementation of the results would facilitate small producers' access to financial services; thus contributing to poverty reduction.