Judicious management of our natural resources is a major pillar of sustainable agriculture, broadly defined. It is, therefore, unfortunate that, until now, research in natural resources management (NRM) has not received sufficient support, or the recognition that other agricultural research topics have had. The Consultative Group on International Agricultural Research (CGIAR) recognized, early on, the complementarities and synergies between NRM research and other themes, notably genetic improvement, and its research Centres have pursued some vigorous programmes for many years. The Technical Advisory Committee (TAC) to the CGIAR was supportive of such programmes from the start and provided advice to the CGIAR System on NRM research strategies. Highlights of such advice over the last 20 years are discussed, including the ecoregional approach, the 1996 study of Priorities and Strategies for Soil and Water Research, and the CGIAR Vision and Strategy of 2000. Recent developments at the Centres leading to the promotion of integrated approaches to NRM (INRM) are described, and a set of priority areas in INRM research for future CGIAR activities is presented.

Shortened fallow periods lead to a decline in crop yields in traditional mixed-food crop fields in southern Cameroon. Farmers use no inputs such as fertilizers for crop production in these field types. Planted fallows with adapted tree species might sustain or increase crop production under those conditions, as found in other parts of Africa. Two-year-old calliandra tree fallows, with trees planted in alleys, clusters, equidistantly or around plot borders, were compared with two- and four to five-year-old natural fallows for their effect on groundnut, maize and cassava yields on eighteen farmer fields in southern Cameroon. Trial fields covered a wide range of soil and environmental conditions. The tree fallows had no significant effects on the yields of maize and cassava with the exception of the border planting, in which trees were not rigorously pruned back prior to cropping. Here, cassava tuber yields were reduced. Cassava tuber yields declined generally with decreasing planting distance to trees. The yield of groundnut, the most important crop in this field type, was adversely affected. However, maize grain yields were positively related to biomass produced by calliandra trees, notably on soils with pH over about 5.3. The data indicated that yields of all crops could be increased with higher plant densities, irrespective of fallow type. The tree fallows, with the exception of border planting, showed less adaptability than four to five-year-old natural fallows to sites with low crop yields, whereas no difference compared with two-year-old natural fallows was found.

To understand the impact of residue and tillage management on crop productivity of rice and barley (including grain yield) and soil fertility in a tropical dryland agro-ecosystem, the following six treatments were established: (a) conventional tillage and residue removed (CT−R); (b) conventional tillage and residue retained (CT+R); (c) minimum tillage and residue removed (MT−R); (d) minimum tillage and residue retained (MT+R); (e) zero tillage and residue removed (ZT−R); and (f) zero tillage and residue retained (ZT+R). Minimum total net productivity (TNP) in both barley (the winter season crop) and rice (the rainy season crop) was recorded in the ZT−R treatment and the maximum in the MT+R treatment. In these crops, 83–88% of TNP was represented by above-ground net productivity (ANP). A reduction in tillage, from conventional to zero, especially when combined with residue retention, reduced the crop TNP/weed TNP ratio, reflecting the abundance of weeds with zero tillage. In both crops, minimum tillage increased grain yields, compared with conventional, but zero tillage resulted in reduced yields. Soil microbial biomass carbon (MBC) and nitrogen (MBN) and N-mineralization rates were higher in rice than in barley, and were maximum in the MT+R treatment. Crop productivity and grain yield in different treatments were positively correlated with MBC, MBN, N-mineralization rate and available-N. Strong negative correlations between crop productivity and grain yield with weed TNP and N-uptake were recorded. This study suggests that retention of a small fraction (one-third) of above-ground biomass from the previous crop and its incorporation in the soil through minimum tillage enhances the crop productivity and grain yield of succeeding crops and promotes soil fertility in a cereal based tropical dryland agro-ecosystem.

Soil phosphorus fractions, maize and cassava P uptakes in natural regrowth fallow and Pueraria phaseoloides fallow systems, as affected by the duration (1–3 years) of fallow, were assessed in 1998 and 1999 in a long-term fallow management trial, established in 1989 at the International Institute of Tropical Agriculture, Ibadan, Nigeria. After one and three years fallow respectively, the mean P accumulation in fallow vegetation was 4.2 and 9.4 kg ha−1 in 1998, and 8.9 and 12.1 kg ha−1 in 1999 for natural fallow. For Pueraria, the corresponding values were 6.3 and 7.4 kg ha−1 in 1998 and 6.1 and 13.5 kg ha−1 in 1999. The positive effects of fallow lengths on dry matter and P accumulation were not as pronounced under Pueraria. Continuous cropping reduced soil pH and depleted Olsen P more than crop-fallow rotations under both fallow systems. An increase in fallow duration enhanced the availability of resin P, NaHCO3-, and NaOH- extractable P fractions, especially under natural fallow. Continuous cropping in a natural fallow system resulted in lower resin inorganic P (Pi) and NaHCO3-extractable inorganic and organic P (Po) fractions, and higher refractory P fractions (concentrated HCl-extractable Pi and Po and residual P) as compared to crop-fallow rotations. Although continuous cropping in the Pueraria system resulted in higher refractory P fractions than the crop-fallow rotations, the effects of fallow length on P fractions were not consistent or pronounced. Maize and cassava P uptakes increased with fallow duration. Phosphorous in maize grain was positively correlated with fallow vegetation P accumulation. Total P uptake by maize was positively but weakly correlated with NaHCO3-Po, Olsen P and fallow vegetation P accumulation.

In eastern Zambia, farmers prefer direct seeded improved fallow species like Tephrosia vogelii to other species because of the reduced labour requirement for establishment. Since relying on two locally available provenances of T. vogelii could become ecologically unsustainable, due to pest and disease outbreaks, a wider range of usable direct seeded Tephrosia provenance material is desirable. A farmer participatory evaluation of Tephrosia candida provenances for their effectiveness as improved fallows and provision of construction materials and fuel wood was conducted. Using an indigenous board game called bao, 68 farmers rated the four provenances and five control species/provenances for these benefits. The T. candida provenances were rated high for soil fertility improvement. First year T. candida post-fallow performance showed that maize yields could be obtained that were comparable with yields following Tephrosia vogelii provenances. According to farmers' ratings, T. candida provenances could provide better fuel wood and light construction materials than T. vogelii, although they were not considered to be as good as Sesbania sesban and Senna siamea. Tephrosia candida provenances have the potential to be tested by many farmers on-farm. Farmers' methods of screening trees can complement scientific predictors, and their early integration into the selection and design of agro-forestry systems can improve relevance and adoption.

Potatoes are an important cash crop for small-scale producers worldwide. The move away from subsistence to commercialized farming, combined with the rapid growth in demand for processed agricultural products in developing countries, implies that small-scale farmers and researchers alike must begin to respond to these market changes and consider post-harvest treatment as a critical aspect of the potato farming system. This paper presents and assesses a low cost potato-grading machine that was designed explicitly to enable small-scale potato growers to sort tubers by size for supply to commercial processors. The results of ten experiments reveal that the machine achieves an accuracy of sort similar to commercially available graders. The machine, which uses parallel conical rollers, has the capacity to grade different tuber shapes and to adjust sorting classes, making it suitable for locations with high potato diversity. Its relatively low cost suggests that an improved and adapted version of this machine might enhance market integration of small-scale potato producers not only in Peru, but in other developing countries as well.

In Mozambique, the sweet potato (Ipomoea batatas) is grown both as a leafy vegetable, the terminal shoots or vines being progressively harvested during the season, and as a root crop. It is produced in the dry season, in areas with a high water table or with irrigation, and in the rainy season. This paper reports the results of measurements made during the 1995 dry season to evaluate the effects of water availability and the frequency of vine harvesting on the processes of radiation interception, dry matter production and partitioning. An irrigated crop (cv. TIS 2534) with a single, end of season vine-harvest (H1) intercepted 71% of the total solar radiation, reducing to 52% with weekly (H4) harvests. The corresponding values for a rain-fed crop were 33 and 20% respectively. When the leaf area index (L) exceeded 3–4, virtually full interception of photosynthetically active radiation (PAR) occurred, except when frequent vine-harvests modified the architecture of the leaf canopy. The extinction coefficient (k, PAR) was highly variable: at a given leaf area, interception was greatest earlier in the season, with infrequent harvests and with irrigation. Total dry matter production (vines, foliage and storage roots), over a period of 161 days, reached 23 t ha−1 in the single harvest (H1), irrigated crop, less when the vines were harvested weekly (H4). The corresponding value for the rain-fed crop, which was able to resume active growth immediately after a rainfall event, following a prolonged dry period, was 7 t ha−1. The efficiency of conversion of PAR into dry matter (εs) was 2.74 g MJ−1 in the irrigated treatments combined, decreasing to around 73% of this without irrigation. The value of εs appeared not to be influenced by dry air or high temperatures. The reduction in εs due to drought was less than the corresponding reduction in intercepted PAR. Increasing the frequency of vine harvests increased the partitioning of assimilates to vines. Irrigation had similar effects early in the season, but afterwards it encouraged preferential growth of the storage roots. The end of season harvest index (h) for vines increased with harvesting frequency from, for example, 0.08 (H1) to 0.38 (H4) in the irrigated crop. By comparison, irrigation increased ‘h’ for storage roots from 0.24 (rain-fed) to 0.57 in the H1 treatment. The total harvest index (vines plus roots) increased with the number of vine harvests from 0.32 (H1) to 0.55 (H4) in the rain-fed crop, and from 0.62 (H1) to 0.72 (H4) in the irrigated crop. Clearly, ‘h’ is not a conservative parameter in the sweet potato, but is sensitive to crop management practices and to soil water availability.

Factorial experiments are described and the importance of interactions emphasized with suggestions about how they should be interpreted. A distinction is made between factors introduced only to see if they will provoke an interaction and those actually under study. Split-plot designs receive special attention. Factorial experiments often involve a large number of treatments and ordinary block designs may be ineffective in controlling environmental variation. If factors have few levels, as in exploratory experiments, the usual device for reducing block size is confounding, which is explained along with single-replicate experiments, partial replication and hidden replication. Alternatively, non-orthogonal designs and analysis of data by nearest-neighbour and spatial methods might prove useful. The need for randomization and the role of significance are discussed. It is pointed out that interactions can sometimes be avoided by transformation of the variate.