Herding and farming environments in sub-Saharan Africa are described briefly. The relative importance of crops and livestock is described for each environment. A matrix of crop-livestock interactions is defined for each environment and the reasons for each type of relation are noted. Farming systems are defined as segregated or completely integrated, according to the degree of crop-livestock integration on the farm. Points at which farming systems research (FSR) might be undertaken to improve interactions between crops and livestock are noted for both systems.

The conservation of genetic resources is discussed in relation to the stages in the evolution of crop plants. Strategies for the collection and conservation of wild species, landraces and advanced cultivars are considered in relation to the pratical needs of plant breeders and other users.

Gypsum application and irrigation increased yield in early maturing groundnut genotypes in experiments using line source and conventional irrigation. Response to gypsum varied with genotype; with some gypsum increased yields at all water application rates, in some genotypes there was no response, while with others gypsum increased yield in drought conditions.

In a separate study of the effect of gypsum and drought on pod initiation and development in three groundnut genotypes, gypsum did not greatly influence pod initiation when adequate irrigation was applied, but was beneficial when water was withheld during pod set and again during pod filling. Crops where the combination of gypsum and genotype were most advantageous in the first drought period subsequently grew more slowly so that there were no final differences in response to gypsum. Cultivar EC 76446(292) had a higher requirement for gypsum and was more susceptible to drought than the other two genotypes.

The seasonal water use patterns and leaf water potential of chillies (Capsicum annum var. annum) grown as a monocrop and as an intercrop between soyabeans (Glycine max) were studied under a weekly and two weekly irrigation regime on a red duplex soil in Northern Victoria, Australia. Irrigation at two weekly intervals resulted in a temporal stratification of water extraction over the growing season, the soyabeans making their maximum demand during the reproductive phase. Senescence of the soyabeans at the R6–R7, growth stage was accompanied by an increase in water uptake by the intercropped chillies. No corresponding increase in water uptake was observed in monocropped chillies irrigated every two weeks or in intercropped or monocropped chillies irrigated weekly. Leaf water potential of the intercropped chillies was, for the most part, greater than that of the corresponding monocrops.

Hypocotyl length in groundnut is a function of sowing depth. In field experiments deep sowing increased the mass of hypocotyl but decreased that of roots, pods and haulm. Few nodules were formed on the hypocotyls of plants from shallow sown seeds (4–5 cm deep). More hypocotyl nodules occurred on Virginia types when deep sown but the number and activity of nodules on the roots decreased. Nodules on the hypocotyl appeared later and fixed less nitrogen than root nodules. Although hypocotyl nodules fixed nitrogen during the later stages of plant growth, this activity could not compensate for the loss in nitrogenase activity due to deeper sowing. Deeper sowing also resulted in decreased pod yields.

Attention is drawn to the large annual fluctuations of clove yields in Zanzibar and Pemba which appear to result from a synchronization of clove production by trees. Linear regression analyses are presented which show that in years of peak production tree yield is often correlated with tree canopy area and that when trees yield in two consecutive years, yield in the second year is usually negatively correlated with yield in the first year; equations including both variables often leave considerable variation in yield unaccounted for. It is suggested that some of the year to year variation in tree yield may result from an environmental stimulus affecting clove bud differentiation.

Field experiments were conducted on two soils (clay loam and sandy loam) with varying texture and water table depth to study the effects of two planting methods (dry seeding and transplanting) on the root growth and grain yield of rainfed lowland rice. About 75% of the roots were found in the top 10 cm soil. Root length density at the flowering stage and grain yield in the sandy loam soil, with a water table deeper than 1.5 m, were less in dry-seeded than in transplanted rice, probably because of the greater moisture stress in dry-seeded rice between late tillering and harvest. In the clay loam soil, with the water table mostly at the soil surface, there was no difference in grain yield between dry-seeded and transplanted rice.

Extracted leaf protein is now used regularly as human food in several places. For preliminary, or domestic, trials the equipment used is inconveniently large. A miniature extraction unit, which does not depend on electricity, is described.

The effect of the field application of chlormequat on the growth, yield and seed oil content of Brassica juncea grown under irrigated and rainfed conditions was evaluated for three consecutive crop seasons. Foliar application at the flower initiation stage increased the number of primary branches, number and length of siliquae, weight of seeds and seed oil content. Chlormequat (50 AS) applied at 80 ml ha−1 increased seed yield by more than 50%. The increase was greater in unirrigated crops.

Potatoes (Solanum tuberosum L.) in central Africa are often grown between two rainy seasons to avoid late blight infection. The influence of planting date and setaria mulch treatment on soil temperature, soil humidity, late blight development, ground cover with green leaves, intercepted radiation and tuber yield of potato was studied in crops grown at 1850 m above sea level in Rwanda.

The planting dates towards the end of the rainy season resulted in the best yields and these were considerably increased by mulching which retained soil moisture and reduced soil temperature so resulting in a better ground cover. Increased yields in the mulched plots seemed to be due to increased radiation interception rather than to better radiation use efficiency.

General mathematical relations between yield parameters, plant populations and rainfall were developed for an indigenous sorghum from the results of 28 population/row spacing trials conducted at four sites over five seasons. Populations maximizing yield increased from 25 000 to 69 000 plants ha−1 over the rainfall range 200–700 mm (pre-planting to harvest total). Tillering partly compensated for low populations but yields from 10000 plants ha−1 at 300 and 600 mm rainfall were only 80 and 61% of potential maximum, respectively. Row spacing at constant population affected tiller numbers and eventual panicle weights but not panicle numbers, and any yield differences were unrelated to rainfall.

Regional population recommendations for sorghum production in Botswana were calculated by applying long-term meteorological records to an experimentally developed model linking sorghum yield to plant population and rainfall. The basic recommendation at any place was taken to be that population maximizing local long-term yield means while keeping the risk of crop failure to one year in ten or less. Values ranged from 25 000 to nearly 70 000 plants ha−1, according to rainfall amount and reliability. The modifications of these basic recommendations necessary to allow for differences in standards of husbandry and for sorghum variety are discussed.