The effects of temperature and photoperiod on times from sowing to flowering (f) were investigated in medium-and late-maturing pigeonpea (Cajanus cajan). Twelve genotypes were sown in two seasons at seven sites in Kenya, covering latitudes 0–4°S and a wide range of altitudes (50–2000 m), as well as under polythene enclosures constructed at six sites to create warmer temperature regimes (a total of 26 environments). The same genotypes were also sown at monthly intervals and in an artificially extended photoperiod (in the open as well as under polythene) created by incandescent lamps suspended above the plots at Katumani (1°30′S).

Times from sowing to flowering varied from 70 to more than 300 days and were associated with variations in mean pre-flowering values of temperature and photoperiod which ranged from 15.2° to 32.7°C and from 12.6 to 15.0 h d−1. Genotypic variation in f in the most inductive regimes (a mean pre-flowering temperature of 24.3°C for the medium- and 20.8°C for the late-maturing genotypes, combined with a mean pre-flowering photoperiod of 12.6 and 12.8 h d−1) ranged from 70 and 76 days and from 85 to 112 days, respectively. There were no photoperiodic effects on f over the range from 12.6 to 13.1 h d−1, but the artificially extended day delayed flowering, especially in the late-maturing genotypes.

The relation between l/f and the mean pre-flowering temperature was linear below and above an optimum temperature, To. The genotype-specific parameters derived from these thermal linear rate models based on flowering responses in 26 environments closely predicted l/f and therefore f in an independent sequence of monthly sowings. It was thus responses to temperature below and above To and not responses to daylength which modulated flowering throughout the wide range of natural environments tested within this vast country, even in the late-maturing and most photoperiod-sensitive genotypes.

Sorghum landraces from contrasting districts of northern Nigeria were grown during the 1990 rainy season to describe their physiological and morphological features. Changes in their dry matter production and yield could be predicted from thermal time to flowering (based on response to a fixed daylength) and partitioning indices. Many other morphological and physiological features, such as light-extinction coefficient, light-use efficiency, plant height, leaf area and leaf number, were either stable or varied systematically with time to flowering. Grain yields of early maturing lines were limited by low light interception from flowering to physiological maturity and those of later maturing lines by highly site-specific drought stress.

The grass Leptocloa fusca is very useful on salt-affected soils as it can tolerate extremely saline and alkaline conditions, Since its growth is not affected by gypsum application, planting with Leptochloa is a good biological method for the reclamation of alkaline soils. It is also well adapted to the waterlogging encountered on saline and alkaline soils, and improves the soil's physical, chemical and biological properties so that within two or three years many commercial and forage crops can be grown. Leptochloa excretes salts through specialized glands and is therefore reasonably palatable to farm animals. Because of its vigorous growth on alkaline soils it does not allow satisfactory growth of companion trees, especially in the initial years of soil reclamation.

The composition and control of the grass weed flora of India and Pakistan are discussed, and the losses caused by the weeds are reviewed. The most important perennial weeds are Cynodon dactylon and Sorghum halepense and the most important annual weeds Echinochloa crusgalli and E. colona in the summer (kharif) crops and wild oats (Avena spp.) and Phalaris minor in the winter (rabi) crops. Losses caused by these weeds can be considerably reduced by the use of selective herbicides but these may be too expensive for many farmers. The problems and control of Phalaris minor are considered in detail.

The role of straw yield and quality in farmers' decisions about the adoption of pearl millet cultivars in western Rajasthan was examined using farm household surveys. Information on the relative importance of grain versus straw yield, the perceived risks associated with the use of new cultivars under variable climatic conditions and perceptions of straw quality indicated that improved cultivars of pearl millet were not adopted primarily because of poor grain yield in years with low rainfall, though poor straw yield in such years was also important. These results are particularly pertinent in the light of farmers' perceptions of the likelihood of experiencing drought or low rainfall. If new cultivars of pearl millet are to replace the traditional ones, they must perform better under conditions of limited rainfall.

Soil fertility decline and fallow effects were studied in Ferralsol-Acrisol catenas of plantations of sisal (Agave sisalana) in north-east Tanzania. The fertility of Ferralsols that had been subject to continuous sisal cultivation in the absence of fertilizers was extremely low but that of Ferralsols that had been under 18 years of bush fallow or under secondary forest was slightly better. Acrisols that had been under continuous sisal cultivation were less depleted than the Ferralsols because of greater intrinsic fertility. A comparison of soil analytical data from the 1950s and 1960s with recent data from the same sisal fields showed that the topsoil pH of the Ferralsols had decreased by 1.5 (r2 = 0.807) and that of the Acrisols by 1.2 (r2 = 0.494) under continuous sisal cultivation. Thus there had been a serious decline in soil fertility under sisal cultivation, and this decline was not adequately reversed by fallowing.

The effect of spatial variation on experimental error variance and the significance of differences between genetic entries was evaluated in five performance trials. A significant portion of spatial variation could be detected in all the experiments investigated and various neighbour plot residuals were applied to adjust for local field trends. Neighbour plot adjustment was clearly more efficient than the randomized complete block design in reducing error variance and in detecting significant differences between entries. It was also more efficient than lattice designs in trial fields exhibiting short distance trends, which could not be covered efficiently by incomplete blocks. In most experiments with long and narrow plots, longitudinal adjustment using two neighbours at each side of a test plot was superior to adjustment by only one ‘nearest’ neighbour.

Surveys of the relationship between pests and diseases and crop sustainability in several areas of Kenya, Nigeria, and other regions of Africa indicate that the production of numerous crops has declined sharply as a result of major pest and disease outbreaks, and others are threatened with major decline because of a surge in virulence of an endemic pest or disease, the introduction of a virulent exotic pest or pathogen, or because a system of control used previously has collapsed. Many of the crops that have declined were already experiencing reduced economic demand. In other cases, crop sustainability has been preserved by vigorous farmer responses or by the intervention of national and international research institutions. Most pests and pathogens, however, remain within tolerable bounds most of the time, though this often requires the use of chemical or cultural controls, or the availability of adequate land to compensate for losses. As land availability declines, more strenuous management efforts may be needed to sustain productivity.

The competition for fertilizer nitrogen between sugarcane and a companion crop (maize or potato) grown between the sugarcane rows was studied using nitrogen-15 labelled nitrogen in three field experiments in Mauritius. The effect of the timing of nitrogen application on nitrogen recovery by sugarcane was also investigated. Not more than 15 kg ha−1 of the 120 kg ha−1 nitrogen applied to the sugarcane was taken up by the companion crop but this was compensated for by the uptake of 8 kg ha−1 nitrogen applied to maize or potato. The present recommendations for nitrogen fertilization of pure stand sugarcane were found to be applicable to sugarcane intercropped with non-leguminous food crops. Though fertilizer nitrogen uptake by sugarcane was increased by delaying nitrogen application until after the harvest of the maize and potato, this was not accompanied by an increase in sugarcane yields.

On-farm experiments were used for assessing the processing and food value of four promising sorghum cultivars and the willingness of farmers to pay for the herbicide 2,4-D for controlling the parasitic weed Striga hermonlhica in coarse grains. The active participation of farmers in these experiments was useful in determining the conditions under which the improved sorghum cultivars and the herbicide would be most likely to he adopted by smallholder farmers in southern Mali.

An experiment to investigate the effect of planting pattern, time of planting the cowpea, and insecticide application on the yield of intercropped cotton and cowpea showed that insecticide applied to the cotton controlled cowpea pests best when the crops were planted in alternating single rows. I n wet seasons, delayed planting of the cowpea reduced its yield but its performance was better in alternating single rows than in other planting patterns. In drier seasons, cowpea yield was unaffected by time of planting and yields were similar with alternating single or double rows. Cotton yield was reduced by 47% when the planting pattern was changed to a one cotton:two cowpea row configuration.