In a field experiment, microbiological and chemical amelioration of a highly deteriorated alkaline soil (pH 10.6, exchangeable sodium percentage 95) was undertaken during 1989–92 using two reclamation technologies: namely, growing Karnal grass (Leptochloa fusca) as a first crop in the absence of any amendment (biological reclamation) or applying gypsum as a chemical amendment for different cropping sequences. After ten months average dehydrogenase activity, used as a measure of catabolic activity of micro-organisms under anaerobic conditions, was 49.5 μg triphenylformazan (TPF) per g soil in treatments where the soil was reclaimed by growing Karnal grass, compared with 26.8 μg TPF by applying 50% of the estimated requirement of gypsum. After three years of reclamation, average dehydrogenase activity was 118.7 μg TPF per g soil with Karnal grass and 96.1 μg TPF with gypsum. After three years average microbial biomass carbon was 28% greater in the gypsum treatments (206.5 mg per kg soil) compared with Karnal grass treatments (161.7 mg). However, at this stage the ratio between average dehydrogenase activity and microbial biomass carbon was greater with Karnal grass (0.77) than in gypsum treatments (0.47). The microbiological properties changed more than the chemical properties of alkali soil as the time period advanced. After three years of reclamation, the average pH of the alkali soil dropped from 10.6 to 9.45 and the average exchangeable sodium percentage was reduced from 95 to 47.5. The increase in total organic carbon was 64% and in available nitrogen about 38% compared with the original soil. Higher microbial biomass carbon in gypsum treatments corresponded with lower exchangeable sodium percentage, higher availability of nitrogen and potassium and greater total yield of sensitive winter clovers.

Advanced genotypes of common bean (Phaseolus vulgaris L.) were tested on-station and on-farm to evaluate their performance and acceptability under farmers' circumstances. Farmers were given seed of the genotypes to plant a minimum area of 100 m2 to be grown as a sole crop or intercropped with sorghum according to their preference. Yield, seed colour, taste, maturity time, disease reaction, growth habit and many other bean characteristics were assessed by farmers. The promising genotypes were higher yielding than the farmers' genotypes in most testing sites. Pairwise ranking of the genotypes showed that in export-oriented western Hararghe, white pea bean (navy bean) type Awash-1 and food bean G-2816 were preferred. In eastern Hararghe GLP ×92 and G-2816 were highly preferred but white pea beans were not preferred. Genotypes in advanced observation nurseries were also evaluated by the farmers and some of the genotypes have been promoted for further testing in yield trials. Selection criteria have been proposed by the farmers in each region. Future breeding and selection of genotypes will incorporate the proposed selection criteria and priorities of the two regions.

A series of grazed and cut experiments were conducted in Bali and North Sulawesi, Indonesia, between 1992 and 1994, to evaluate a range of selected forages for both intensive and extensive ruminant production under mature, tall coconuts. Yield, botanical composition and persistence of grass–legume mixtures were assessed under regular grazing and cutting regimes.

In North Sulawesi, total dry matter yields of most grass–legume treatments were relatively stable over the two-year period but yields of legume components within treatments declined sharply. In Bali, the grazed plots maintained stable yield and legume components over the experimental period but yield and legume components of cut plots declined sharply. Paspalum malacophyllum and Setaria sphacelata cv. Splenda were productive and persistent grasses and formed weed-free swards but did not combine well with prostrate legumes. The erect growth habit of these grasses may limit their suitability in coconut production systems by impeding nut collection. The prostrate species Brachiaria humidicola and Stenotaphrum secundatum were slightly less productive but equally weed-free and persistent, and established rapidly from cuttings. S. secundatum also formed stable associations with prostrate legumes under grazing. Arachis pintoi cv. Amarillo and A. glabrata CPI 93483 were the most persistent of a range of prostrate legumes. The tree legume, Calliandra calothyrsus, showed promise as a productive cut-and-carry species.

This study applies systems simulation to characterize determinants of sustainability of a hillside farm in the upper Cauca River watershed of Colombia. A farm model linked to process-level crop models was used to simulate a set of farm scenarios. Susta inability, expressed as the probability of farm continuation, was estimated for each scenario based on replicated simulation with stochastic weather and price inputs. Hypotheses about determinants of sustainability were tested by sensitivity analysis. Res ults identify cropping system, area under cultivation, consumption requirements and crop prices as important determinants of sustainability. The study highlights the impacts of price variability and spatial diversification on farm risk and the results sug gest ways to enhance the sustainability of the farm. The methodology was useful for integrating the diverse biophysical and economic factors that affect farm sustainability.

A two-year study considered the effect of pod harvesting intervals of 3, 5, 7, 9, 11 or 20 weeks after pod set (WAPS) on the relationship between the physical condition of the pods, seed maturation and quality in fluted pumpkin (Telfairia occidentalis). Pod characteristics, such as the intensity of bloom wax on the pod and stiffness of the pod ridges, showed perfect linearity and correlation with the age of the pod, whereas changes in seed quality factors – germination, dry matter, oil and protein concentrations were curvilinearly associated with the time of pod harvest. Thus, seed quality factors increased steadily up to nine WAPS then declined. The point at nine WAPS defined as physiological maturity for fluted pumpkin seed and approximate visible indications of this point in the field were fairly stiff pod ridge condition and fading of bloom wax.

Common bean (Phaseolus vulgaris L.) germplasm was screened for resistance to charcoal rot (Macrophomina phaseolina) under field conditions at Kiboko and Katumani, eastern Kenya. Of the 313 bean accessions evaluated, 50 lines were resistant and six were tolerant to M. phaseolina, the charcoal rot incidence was less than 25% and between 25% and 50% for the resistant and tolerant lines respectively. Yields ranged from 135 to 1051 kg ha−1 compared with 55 kg for the susceptible control A464. Time to maturity did not seem to influence or affect the susceptibility or resistance to M. phaseolina of the various bean accessions.

The introduction of appropriate crop rotations is known to be beneficial in many farming systems. One feature of rotations is that it takes a valuable length of time for the advantage of the rotation to take effect. In long-term rotation trials, the observations from the same plot over years are correlated; ignoring such correlations may affect the precision of the estimates of rotation effects. We examined five covariance structures between the plot errors over time to assess the effect of correlations on the standard errors of rotation means and rotation x cycle combination (interaction) means on wheat yields using eight years of data from six two-phase rotations with wheat. Based on wheat yield data from the four cycles of the rotations considered, the compound symmetry covariance structure (constant correlation) between plot errors arising over alternate years gave more efficient estimates of rotation means compared with the other four covariance structures.

Comparisons of single cropping, double cropping and relay cropping of soyabeans (Glycine max) with wheat (Triticum aestivum) were made at two sites in Arkansas over a two-year period. The comparisons were made using both soyabean blends and their component cultivars. In relay cropping the yields of pure lines of soyabeans were reduced by 17% compared with single-cropped soyabeans, but the yield of wheat in relay cropping was 15% less than in double cropping. Double cropping reduced the branch number, plant height, node number and leaf area of soyabeans compared with single cropping, but relay cropping reduced only node number and leaf area. Both double cropping and relay cropping gave greater land equivalent ratio (LER) values than single cropping, and double cropping gave greater LER values than relay cropping. Net returns were lowest with single cropping and greatest with double cropping. Soyabean blends gave yields similar to the mean of the component genotypes in all three cropping systems, and the net returns of blends were similar to those of the means of component genotypes.

Erosion trials under natural rainfall were conducted at two locations in the Andean hillsides of south-west Colombia on moderate slopes. Sediment-bound losses of seven cassava (Manihot esculenta) cropping systems per cropping period ranged from 26 to 1726 kg ha−1 for organic matter, 0.9 to 65.5 kg ha−1 for total nitrogen, 0.03 to 2.1 kg ha−1 for exchangeable magnesium, 0.04 to 2.8 kg ha−1 forexchangeable potassium and 0.004 to 0.8 kg ha−1 for Bray-II phosphorus. Runoff water contributed substantially to total available potassium, magnesium and phosphorus losses. Concentrations of soluble phosphorus in runoff water from heavily eroded continuously clean-tilled fallow plots were slightly higher than from cropped and fertilized plots. Enrichment ratios for sediments from these highly aggregated Inceptisols were low. Sand-sized particles were selectively removed by water erosion. The higher proportion of nutrient losses with run-off water in our study stressed the importance of runoff control along with attempts to control soil losses and thus diminish both nutrient losses and surface water eutrophication.

Experimental Agriculture (1997), 33, 367–383.

An incorrect version of Fig. 1(b) was published on page 373 of the above mentioned paper. The correct version appears below. The publisher apologises for this error.