The development of successful improved sorghum varieties for the Western Africa requires the incorporation of farmer’s perceptions and desires into the end product. Failure to do this in the past probably explains the low rate of adoption of improved varieties in Niger.

Participatory rural appraisal (PRA) tools, such as focus group discussions and survey questionnaires, were used in this study to collect data from farmers on their sorghum farming systems, their production constraints, and their preferences and their acceptance of new improved sorghum varieties. A survey involving 100 small-scale farmers was conducted in two villages in the region of Maradi. Results in this study revealed that sorghum production systems in Niger are improving even though the practice of agriculture is still largely characterized by marginal growing environments. Results showed that 46% of farmers were growing improved varieties with inorganic fertilizer and fungicides in very low quantities. Relatively high incidences of heat, drought and low soil fertility, spatial variability in rainfall patterns, low use of external inputs such as improved varieties, and the use of traditional agricultural implements characterized the growing environments. Major production constraints include lack of modern agricultural equipment, parasitic pests and diseases (Striga, sorghum midge, mildew, long smut), lack of education, lack of improved varieties, poor soils, the weather (drought, wind), long maturity period of landraces, low grain yield potential, and lack of sufficient arable land. Earliness and high yield of improved varieties were the most important criteria for farmers to choose a new variety, but they indicated they would not totally reject their local varieties because of social considerations. Cultivar improvement should therefore target characteristics of local varieties in the creation of new ones. Early maturing varieties with high yield potential, resistance to drought, sorghum midge, downy mildew and long smut would be welcomed by farmers.

Cytoplasmic male sterility system has been widely used for increasing sorghum yield through hybrid exploitation. To effectively use male–sterility inducing cytoplasm, it is necessary to identify restorers and lines that are suitable for conversion to male sterility. The present study was undertaken to determine the fertility/sterility reactions of a subset of entries from a total of 1019 accessions of two collections (1976 and 2003) of landrace sorghums crossed onto two CMS lines NE223A and TX623A. The male-sterile lines were Kafir-milo derivatives and have the same cytosterile A1 mechanism. Three hundred and forty test-crosses were generated and evaluated in two environments in Niger. Studies revealed four F1 test-crosses (NE223A x L95-1, NE223A x L95-3, NE223A x L119-2 and NE223A xL102-1) with complete sterility across the two locations. The corresponding parental lines were therefore recommended for conversion to male sterile. However, forty (40) additional test-crosses showed male sterility at Maradi only and two others at Konni whereas, respectively, 286 and 272 hybrids shed pollen. The pollen parents of the fertile testcrosses were classified as fertility restorers. The female NE223A sterilized more lines (42) than TX623A (8) while, at the same time, there were more sterile crosses within the 2003 collection than within the 1976 collection. According to the classification of sorghum races, the sub-group guinea-margaritiferum had the highest number of lines without restorer genes. Other groups tested were Guinea, guinea-caudatum and durra-caudatum.

The limited availability of improved varieties and the low production potential of landraces have constrained sorghum production in Niger. However, hybrid cultivars have been shown to be more productive than pure line and landrace varieties. Hybrid development can therefore enhance sorghum productivity and food security in this Sahelian country. Thirty nine local landraces were crossed onto two CMS lines in a line x tester fashion in the off-season 2010. The 78 F1s with their 39 parents, the two maintainer lines and two checks were evaluated in a split-plot alpha design with three replications in four localities in 2011. The study was conducted to estimate the combining ability of local landraces, to assess heterosis over the best parent and to evaluate sorghum hybrids across a range of environments for possible hybrid release. Analysis of variance revealed that mean squares due to entries, parents and crosses were highly significant for all the traits. Mean squares due to general combining ability (GCA) and specific combining ability (SCA) were highly significant for grain yield. Overall hybrid mean yield was significantly higher than that of parents and check varieties.Hybrids displayed better parent heterosis of 33 to 97%. Lines L16-2, L279-1, L15-2 and L43-3 of the local landraces were identified as good general combiners for grain yield. Based on SCA effect and grain yield across the tested environments, six hybrids(TX623A x L279-1, NE223A x L15-2, NE223A x L16-2, NE223A x L43-3, NE223A x L150-2 and TX623A x L156-2) displayed good performance across and at some individual environments. Both additive and non-additive gene effects were involved in the variations observed among crosses. It was concluded that high yielding and stable F1 hybrids were obtainable from crosses of local landraces and some improved lines.