Constraints to effective weed management may be the main reason for the small area under minimum tillage (MT) in smallholder farming in southern Africa. The effect of maize residue mulching and intensity of hand hoe weeding on the growth of weeds, cowpea (Vigna unguiculata cv. IT 86D-719) and sorghum (Sorghum bicolor cv. Macia) was investigated in the fifth and sixth years of a conservation agriculture (CA) field experiment at Matopos Research Station (28 0 30.92`E, 20 0 23.32`S). The experiment was a split-plot randomized complete block design with three replications. Tillage was the main plot factor (conventional tillage -mouldboard plough compared against MT systems -ripper tine and planting basins) and maize residue mulch rate (0, 4 and 8 t ha -1 ) the sub-plot factor. Hoe weeding was done either four times (high weeding intensity) or twice (low weeding intensity) during the cropping season. Planting and weeding were done at the same time in all treatments. There was markedly greater early season weed growth in MT systems relative to mouldboard plough (MBP) in both crop species. In sorghum, MT (planting basins: 40.3 kg ha -1 ; ripper tine: 34.8 kg ha -1 ) systems had higher cumulative weed biomass measured after planting than MBP (29.9 kg ha -1 ) system. Maize mulching was generally associated with increased mid-to late-season weed growth in the two crops probably due to improved soil moisture conservation during periods of low precipitation. Weed suppression by the maize mulch was observed only in sorghum and limited to early in the cropping season with no effect observed for the remainder of the sorghum rotation phase. The high weeding intensity treatment had lower weed growth in both crops and better sorghum yield than low weeding intensity. The MT systems had poor crop establishment which translated into low yields. Cowpea grain yield obtained from MT systems was less than 300 kg ha -1 compared to 413 kg ha -1 in MBP. The poor sorghum establishment in MT systems translated into low grain yield as

Rainfed smallholder agriculture in semi-arid areas of southern Africa is subject to numerous constraints. These include low rainfall with high spatial and temporal variability, and significant loss of soil water through evaporation. An experiment was established at Matopos Research Station, Zimbabwe, to determine the effect of mulching and minimum tillage on maize (Zea mays L.) yield and soil water content. The experiment was run for two years at two sites: clay (Matopos Research Station fields) and sand (Lucydale fields) soils, in a 7 • 3 factorial combination of mulch rates (0, 0.5, 1, 2, 4, 8 and 10 t ha À1) and tillage methods (planting basins, ripper tine and conventional plough). Each treatment was replicated three times at each site in a split plot design. Maize residue was applied as mulch before tillage operations. Two maize varieties, a hybrid (SC 403) and an open pollinated variety (ZM 421), were planted. Maize yield and soil water content (0-30 and 30-60 cm depth) were measured under each treatment. On both soil types, neither mulching nor tillage method had a significant effect on maize grain yield. Tillage methods significantly influenced stover production with planting basins giving the highest stover yield (1.1 t ha À1) on sandy soil and conventional ploughing giving 3.6 t ha À1 on clay soil during the first season. The three tillage methods had no significant effect on seasonal soil water content, although planting basins collected more rainwater during the first half of the cropping period. Mulching improved soil water content in both soil types with maximum benefits observed at 4 t ha À1 of mulch. We conclude that, in the short term, minimum tillage on its own, or in combination with mulching, performs as well as the farmers' traditional practices of overall ploughing.

Globally, a range of agronomic factors have been reported to have an impact on the performance of conservation agriculture (CA) and often determine its performance in relation to conventional agriculture (CONV). To assess this performance in Zimbabwe, 48 CA experiments were conducted by the International Crops Research Institute for the Semi-Arid Tropics in the semi-arid areas of southern Zimbabwe from 2004 to 2010, to calculate the weighted mean difference (WMD) through meta-analytical methods. The two CA practices, planting basins (Basins) and ripper tillage (Ripper), were compared with CONV. It was hypothesised that CA results improved yield compared with CONV and that the effect of CA practices on yield is affected by soil type, rainfall amount and distribution and selected management practices, which included rates of inorganic fertilisers and manures and mulching. Basins were superior to CONV in 59% of the experiments and the overall effect was significant (p < 0.001). The effect of Ripper was non-significant. The hypothesis that CA practices result in improved maize grain yield over CONV was accepted for Basins. The WMD for experiments conducted on sandy soils was 0.365 t ha −1 for Basins and 0.184 t ha −1 for Ripper, and in both cases was significant (p < 0.05). For clay soils, only the WMD for Basins was significant. A higher rainfall regime (500-830 mm) resulted in a lower WMD for Basins (0.095 t ha −1 ) and Ripper (0.105 t ha −1 ) compared with 0.151 t ha −1 for Basins and 0.110 t ha −1 for Ripper under lower rainfall (320-500 mm). The overall effect of Basins under the higher rainfall regime was not significant. There was better yield performance for Basins when the rainfall was well distributed; the reverse was noted for the Ripper. The application of 10-30 kg ha −1 of N (micro-dose range) resulted in a higher WMD for Basins than zero N application. Without N application, the WMD of Basins was not significant. For zero manure application in Basins, the WMD was 0.043 t ha −1 compared with 0.159 t ha −1 when manure was applied. The application of mulch depressed the WMD in Basins by 44% and Ripper by 89%. The hypothesis that yield performance under CA is influenced by soil type, rainfall amount and distribution, inorganic fertiliser and manure application was accepted.

Next to drought, poor soil fertility is the single biggest cause of hunger in Africa. ICRISATZimbabwe has been working for the past 10 years to encourage small-scale farmers to increase inorganic fertiliser use as the first step towards Africa's own Green Revolution. The program of work is founded on promoting small quantities of inorganic nitrogen (N) fertiliser (micro-dosing) in drought-prone cropping regions. Results from initial on-farm trials showed that smallholder farmers could increase their yields by 30-100% through application of micro doses, as little as 10 kg Nitrogen ha -1 . The question remained whether these results could be replicated across much larger numbers of farmers. Wide scale testing of the micro-dosing (17 kg Nitrogen ha -1 ) concept was initiated in 2003/2004, across multiple locations in southern Zimbabwe through relief and recovery programs. Each year more than 160,000 low resourced households received at least 25 kg of nitrogen fertiliser and a simple flyer in the vernacular explaining how to apply the fertiliser to a cereal crop. This distribution was accompanied by a series of simple paired plot demonstration with or without fertiliser, hosted by farmers selected by the community, where trainings were carried out and detailed labour and crop records were kept. Over a 3 year period more than 2,000 paired-plot trials were established and quality data collected from more than 1,200. In addition, experimentation to derive N response curves of maize (Zea mays L.), sorghum (Sorghum bicolor (L.) Moench) and pearl millet (Pennisetum glaucum (L.) R.Br.) in these environments under farmer management was conducted. The results consistently showed that micro-dosing (17 kg Nitrogen ha -1 ) with nitrogen fertiliser can increase grain yields by 30-50% across a broad spectrum of soil, farmer management and seasonal climate conditions. In order for a household to make a profit, farmers needed to obtain between 4 and 7 kg of grain for every kg of N applied depending on season. In fact farmers commonly obtained 15-45 kg of grain per kg of N input. The result provides strong evidence that lack of N, rather than lack of rainfall, is the primary constraint to cereal crop yields and that

No abstract