Methods of adjustment for two-dimensional spatial heterogeneity of grain yield were investigated for 224 UK cereal trials. The methods used row and column ‘block’ analysis of plot yields and neighbour analysis based on first differences of plot yields. In 36% of trial analyses for block models and 30% for neighbour models the average variance of variety differences was reduced by more than 10% compared with the better of the one-dimensional row or column models. Compared with complete blocks, 2-D block analysis had a mean efficiency of 153% whereas the conventional 1-D block analysis (by rows) had a mean efficiency of 127%. Similarly, 2-D neighbour analysis had a mean efficiency of 159% whereas the 1-D analysis had a mean efficiency of 137%. Recently, general statistical methods have been developed for two-dimensional design and analysis; their wider use should lead to major gains in the precision of variety trials.

Nine experiments were conducted in the south-east of Ireland on three contrasting soils over a 3-year period (1988–90) to test the effect of six cultivars at four rates of nitrogen on the grain yield and quality (ex-farm) of spring-sown malting barley.

Blenheim gave the greatest grain yield and lowest protein on average. Blenheim and Prisma had similar low protein values but Prisma gave significantly lower (P <0·05) mean grain yields than Blenheim at all three sites. Corniche gave significantly lower (P <0·05 to 0·01) grain yields than Blenheim in all nine experiments. Choice of cultivar had a considerable effect (<16%) on the yield and protein content of malting barley but season and soil type had a much larger effect (2–3 times greater).

The application of increasing increments of fertilizer nitrogen generally increased grain protein and screenings significantly (P <0·05 to 0·01) in all nine experiments but it did not always give increased grain yields. The levels at which nitrogen gave significant grain yield increases depended on season and soil type. Season and soil type had a much larger effect on grain yield and quality than nitrogen rate.

It was not possible to relate grain yield to specific climatic factors except on the light textured soil in 1990 when severe moisture stress in the early part of the growing season reduced grain yield by 32·5% compared with the previous 2-year average.

A field experiment at Gudha Experimental Farm, Central Soil Salinity Research Institute, Karnal, India, in 1983/84 evaluated the effect of gypsum (12·5 t/ha, 50% of gypsum requirement of soil), pyrite (10·2 t/ha, equivalent to gypsum on a sulphur basis), farmyard manure (FYM) (30 t/ha), gypsum + FYM, pyrite + FYM and a control, with saline (ECiW 4·0 dS/m) and non-saline (0·4 dS/m) irrigation water on soil properties and yields of rice (Oryza sativa L.) and wheat (Triticum aestivum L.) in a highly sodic soil (pH 10·5, 96% exchangeable sodium). Application of these amendments enhanced the yield of both crops significantly, irrespective of the quality of the irrigation water used. Soil properties and crop yields were improved in the following order: control < FYM < pyrite < gypsum < pyrite + FYM < gypsum + FYM. Irrigation with saline water resulted in significantly higher yields of both crops than irrigation with non-saline water. Decreases in soil pH and exchangeable sodium and increases in exchangeable Ca + Mg and infiltration rate were greater after rice than wheat in the rotation, particularly when non-saline water was used.

In intermediate deepwater conditions (0–50 cm) at Cuttack, India during 1990 and 1991, rice and dhaincha (Sesbania aculeata) were either mixed-sown in different arrangements (parallel lines and mixed broadcasting) in dry soil by the end of May or rice seedlings were transplanted in plots grown with pure dhaincha after the accumulation of water in the field by mid-July. Incorporation of dhaincha in situ after 48 and 54 days of growth added 60·8–65·2 and 72·9–76·9 kg N/ha in the mixed stands compared with 81·3 and 85·1 kg N/ha in the pure stand in 1990 and 1991 respectively. Although the growth of rice when sown simultaneously with dhaincha was affected initially due to the more vigorous growth and shading by the latter, tillering of rice increased after the incorporation of green manure following an increase in inter-row spacings. The performance of the direct-sown and transplanted rice was similar in 1990, when the water level rose relatively slowly, but in 1991 the transplanted crop yielded significantly less due to the sudden accumulation of water to higher depths (48 cm), resulting in greater mortality of seedlings at planting. However, the grain yield of both directsown and transplanted crops increased with dhaincha green manuring compared with the control (no application of dhaincha) and was equal to the application of 40 kg N/ha applied as urea fertilizer. The highest yield was obtained when rice and dhaincha were grown at a 2:1 ratio in 20 cm wide rows and the clonal tillers uprooted from the adjoining rice rows were planted in lines vacated by dhaincha. The yield was, however, equal to that from treatments where rice and dhaincha were grown in alternate rows at 15 cm spacing or mixed broadcasting. Increase in yield under green manuring was due to greater panicle weight, which was probably due to a continued supply of N following decomposition of organic matter added through dhaincha. Therefore, green manuring of direct-sown rice with dhaincha was beneficial for higher crop productivity under excess water conditions.

Yields from five of the plots on the Park Grass Continuous Hay experiment at Rothamsted, started in 1856, were examined to see if any long-term trends could be detected over the last 100 years. Three of the plots examined are unfertilized; two receive inorganic nutrients every year; all are harvested twice a year. In 1959 the harvesting procedure was changed: yields for the periods before and after this change were examined separately and together. On none of the three unfertilized plots was the slope of the regression of total yield (i.e. first and second cutscombined) on time significantly (P < 0·05) different from zero in either the 1891–1958 or 1960–1992 periods. On both the fertilized plots, there were significant declines in yield with time over the 1960–92 period, assuming that the effects of autocorrelation are sufficiently small to be ignored; there were no consistent changes between 1891 and 1958. There were some significant trends on the five plots when the two cuts were considered separately.

A linear regression model was fitted to the data in an attempt to separate the effects of meteorological variables (rainfall and sunshine hours over selected parts of the year) on total yield from possible long-term effects brought about, for example, by the increasing concentration of CO2 in the atmosphere. On four of the five plots this model accounted for between 12 and 21% of the yield variance in the pre-1959 period and between 45 and 63% after 1960. On the fifth plot, which received the highest level of N, the model accounted for 29% of the variance in the first period but only for 16% in the second period. When a linear trend with time was included in the model, this was not significant on any of the plots over the entire 1891–1992 period, although some significant trends appeared when the two periods were considered separately. The model was also fitted with the atmospheric CO2 concentration in place of the linear trend with time: again there were no consistent trends.

Neither changes in the concentration of CO2 in the atmosphere over the last century nor increasing inputs of combined N in rainfall or in dry deposition have had any detectable effects on yield in these plots.

The morphology and position of the shoot apex in field swards was studied in 12 grasses for 12 months, in eight related grasses in 2 years, and in seedlings, young tillers and older tillers of Lolium perenne.

Lolium multiflorum had relatively large numbers of primordia, long apices, which were above ground or near ground level, and a small length of apex per primordium. Festuca arundinacea had few primordia, a large length of apex per primordium, and apices which were well below ground or near ground level. Lolium perenne, Festuca pratensis and Lolium × Festuca hybrids were intermediate between Lolium multiflorum and Festuca arundinacea in numbers of primordia and length of apex per primordium. Festuca rubra and Dactylis glomerata had rather small numbers of primordia, whereas Phleum pratense had relatively large numbers of primordia and of unemerged and emerged leaves.

The numbers of leaf primordia in Lolium perenne increased as the seedling main shoots or the tillers grew older, while the numbers of unemerged leaves per shoot remained more constant.

The morphology, position and protection of shoot apices are discussed in relation to differences between species in rate of establishment and persistence.

The patterns of N uptake and dry matter synthesis by sugarcane (Saccharum hybrid spp.) were studied at four locations in Mauritius with 15N–labelled ammonium sulphate (100 kg N/ha) applied either in a single dressing in September or in two split applications in September and the following February. More than 80% of the total N recovered at harvest (100–120 kgN/ha) was absorbed by the sugarcane during an active uptake period from October to January. Split application prolonged this active N uptake until April only and had no effect on dry matter accumulation. While total Nabsorbed by above-ground sugarcane showed no decline over time, 10–20 kg N/ha of the 15N–labelled N was lost from the green tops even when the N was applied on two occasions. The fertilizer N losses from above-ground sugarcane were, however, not evident when fertilizer N recovery with time was studied by the difference method. In view of the observed losses of fertilizer N from the aerial parts of sugarcane, measurement of fertilizer N recovery at harvest by the N isotope dilution technique underestimates fertilizer N uptake by sugarcane and attributes too large a fraction of N loss to denitrification/volatilization of NH3.

In the first of two series of experiments the effects of barley straw disposal by burning, chopping and spreading or baling and removing on winter oilseed rape were tested after seedbed preparation either by ploughing in the residue or incorporation in the soil by tine cultivation. These treatments were compared in four field experiments on silty clay loam soils at Rothamsted, UK from 1986 to 1989. The winter rape was either sown in late August or mid-September in seedbeds where either none or 50 kg N/ha had been applied.

In the first season, August-sown rape was successfully established after tine cultivation but very dry conditions prevented seedbed preparation after ploughing, consequently all ploughed treatments were sown in September. Continuing dry conditions delayed emergence of the September-sown crop, the resultant small rape seedlings suffered substantial winter kill in some treatments during a period of abnormally low temperatures. Yield from the August-sown treatments was large (c. 40 t/ha) and showed no effect of straw disposal treatment or additional seedbed N. The yield of the September-sown crop was influenced by the amount of winter kill sustained; the smallest yields resulted from tine incorporation, and the largest after ploughing, where they approached those of the early sown crop.

In the following three seasons more plants emerged from September than from August sowings. The application of seedbed N increased the plant population of the August-sown crop where the straw had been burnt. Plant losses over winter ranged from 15 to 20% and were unaffected by straw disposal treatment. There were significant differences in yield resulting from season and sowing date. Incorporating chopped straw by tine cultivation significantly decreased yield which, coupled with a lower oil content in the September-sown crop, gave a significantly smaller oil yield.

In a second series of field experiments from 1987 to 1989, the effects of improving the timeliness of rape establishment after winter wheat by broadcasting rape seed into standing wheat was compared with conventional sowing after preparing a post-harvest seedbed. After cereal harvest, straw was disposed of either by baling and removing or chopping and spreading over the rape seed. The application of 50 kg N/ha to the stubble or seedbed was also tested, as was the effect of increasing the seed rate from 8 to 16 kg/ha in two seasons.

There were always fewer plants established from broadcasting than from drilling. Generally there were fewer winter losses from broadcast seed than from drilled. Applying N to seedbed or stubble had no effect on plant population or survival except where 16 kg/ha seed was sown in 1989 and increased yield in two of the three seasons.

In spite of a smaller plant population from broadcast seed, yields were often larger from broadcast than drilled treatments. On average broadcasting the seed and baling the straw gave the largest yield although this was significant only in 1989.

Effects of agronomic factors on yield of winter oilseed rape were examined in multifactorial experiments, with a single-low (low erucic acid, high glucosinolate) cultivar (Bienvenu) in 1984/85 to 1987/88 and a double-low (low erucic acid, low glucosinolate) cultivar (Ariana) in 1986/87 to 1988/89. Establishment of September-sown plots was better than that of August-sown plots in 1984, 1987 and 1988, but not in 1985 when September was very dry. The total dry matter (DM) (kg/ha) of August-sown plots was much greater than that of September-sown plots throughout autumn and winter but by harvest the plant DM of late-sown plots was slightly greater. Combine-harvested yields of late-sown plots were greater than those of early-sown plots in 1984/85 (cv. Bienvenu) but not in 1988/89 (cv. Ariana); effects of sowing date on yields were smaller in other seasons, although late-sown plots generally had larger yields. Insecticide treatments increased combine-harvested yields in 1987/88 (cv. Bienvenu) and 1988/89 (cv. Ariana) and increased hand-harvested yields in 1984/85, by increasing the number of plants/m2. Fungicide spray treatments increased yields in all years except 1989, by increasing 1000-seed weights. Application of growth regulator sprays increased yields in 1986/87 to 1988/89; effects on cv. Ariana were associated with increased numbers of pods/plant. There were no consistent effects of spring nitrogen rates or timing on yield or yield components.

Field experiments were conducted during the two monsoon seasons of 1985 and 1986 at Kharagpur, West Bengal, to investigate the effect of pigeonpea cultivars and their planting pattern on the performance of a groundnut–pigeonpea intercrop. In intercropping the crops were also sown at 100% population of their sole crops. Groundnut plant growth and development was affected by intercropping with pigeonpea and the decline in pod yield ranged from 7 to 50% of the sole crop yield. The compact, short-statured and short duration pigeonpea cultivar, ICPL87, intercepted only 50% of the incoming solar radiation as against the 70% intercepted by LRG30 at the peak vegetative stage. The significantly higher groundnut yield (48%) obtained from the intercrop with the former pigeonpea cultivar as compared to the latter justifies the belief that the performance of the groundnut was light-related. This was further confirmed when the light interception by pigeonpea was reduced by increasing the row spacing from 90 to 150 cm and there was a consequent increase in groundnut yield. The cultivar and row arrangement of pigeonpea that allowed equal utilization of light by both crops gave increased productivity.

Studies were conducted to evaluate amounts of fatty acids in the oil as possible indicators of crop maturity in the groundnut, cultivar NC2, grown in St Vincent, Eastern Caribbean. Percentage fatty acids in oil of the composite samples of seeds collected between 99 and 141 days after planting for a period of 3 years (1985–87) was determined by gas-liquid chromatography. Results indicated variation in the amounts of individual fatty acids with crop maturity. There were definite trends, however, with respect to oleic and linoleic acids. As seeds matured, percentage oleic acid increased and percentage linoleic acid decreased, and consequently, the oleic:linoleic acid ratio in oil of the composite seeds increased significantly. The percentage oleic acid peaked significantly at a value of 55.8 ± 0.59 at optimum harvest dates. The results clearly suggested therefore that percentage oleic acid in oil could be used as a reliable reaping index for groundnut.

Water use efficiency (WUE) was measured on fourteen genotypes of groundnut (Arachis hypogaea L.) grown in containers under adequately irrigated and water-limited conditions. The genotypes used similar amounts of water but produced different quantities of dry matter. WUE accounted for > 92% of the variation in dry matter production under both irrigated and water-limited conditions. There was a significant increase in WUE under water-limited conditions. Four genotypes selected from the container experiment as having either a high or a low WUE under non-limited or limited water input conditions were further tested under prolonged water deficit conditions in a field experiment. WUE varied significantly between genotypes and there was a positive correlation between WUE and the quantity of dry matter produced by the genotypes. The results suggested that, in three out of four genotypes, the WUE measured in the container experiment was positively correlated with the WUE estimated under field conditions.

This study was initiated to evaluate the short-term (< 3 years) ability of some selected acid-soil adapted hedgerow species when planted in alley cropping systems to improve soil properties in a newly-cleared Ultisol (Typic Kandiudult) of southern Cameroon, 1990–92. The hedgerow species selected were Senna (Senna spectabilis), Flemingia (Flemingia congesta)and Acioa (Acioa barterii). A non-alley-cropped control was also included in the trial. The greatest quantities of prunings, and hence, mulch were produced by Senna and Flemingia. Exchangeable Ca, effective CEC and water infiltration were greatly increased in the alleys of plots where either Flemingia or Senna had been planted within 2·5 years of hedgerow establishment. The large amounts of mulch produced by Senna and Flemingia did, however, result in soil temperatures greater than those of the control or Acioa plots c. 1 year after application of the prunings as mulch. Root growth of Senna in the subsoil was significantly greater than that of either Acioa of Flemingia, but that of cassava was reduced by alley cropping with all three hedgerow species. Compared to the control or alley cropping with Acioa, maize and cassava yields were greater when alley cropped with either Flemingia or Senna hedgerows.

Nitrogen nutrition of two succeeding wheat crops was studied after ploughing of grassland in July 1987 on a clay soil at ADAS Drayton. The four plots of grassland had received 100, 250, 450 and 750 kg N/ha per year for 4 years from 1984 and were grazed by beef cattle at stocking densities which varied according to grass growth.

Determinations of soil mineral N taken to 60 cm every 3 weeks from July to the following May were particularly variable. However, in the first 2 years after ploughing the means of the series of mineral N measurements were directly proportional to the amounts of fertilizer N applied to the grass.

N offtake in winter wheat grain without fertilizer N was directly proportional to fertilizer N applied to grass but this had little effect on maximum grain yields. Large soil N supplies did not necessarily predispose the wheat crops to large grain N concentrations because fertilizer N caused grain N offtake to reach a similar maximum, irrespective of previous grass N.

Optimum amounts of fertilizer N for the wheat were 188, 147, 87 and nil kg/ha in 1988 and 152,130, 89 and 25 kg/ha in 1989 after 100, 250, 450 and 750 kg N/ha per year applied to the grass. Soil N supply as indicated by both the amount of fertilizer applied to grass and means of mineral N measurements accounted for almost all of this variation. Mean soil mineral N over winter was no better as an indicator of soil N supply than the amount of N applied to the grass. However, before adopting N applied to grass as a more general index of N supply, it would need to be adjusted for variation in N removed and lost during grass growth; these were controlled in this experiment.

This paper reports genetic differences between 20 Mashona bulls from smallholder farms, 14 from Makoholi Research Station, three from commercial farms and two reference sires, in Zimbabwe. An animal model, fitted through the derivative free Restricted Maximum Likelihood program (DFREML), was used to analyse 707 birth weight and 623 weaning weight records. Both the mean breeding values of the reference sires and the ranch bulls were significantly (P < 0·05) higher than the mean breeding values of smallholder bulls for birth weight (by 1·6 and 1·7 kg, respectively) and weaning weight (by 10·4 and 11·0 kg, respectively). However, seven of the 20 bulls from smallholder farms were among the top 45% of bulls with high breeding values for both birth and weaning weight. The mean breeding value of bulls from Makoholi was not different from any of the other groups, for all traits studied. It was proposed that future efforts to form nucleus herds to improve the weaning weights of indigenous cattle should include the genetic pool in smallholder farms.

Two experiments of randomized block design were carried out at the Agricultural Research Institute of Northern Ireland in 1991/92 involving groups of 20 pigs fed ad libitum from single-space feeders with built-in water supply from 37 kg to slaughter at 91 kg liveweight.Treatments were replicated 16 and 6 times respectively in Experiments 1 and 2. In the first experiment, one feeder perpen was compared with two feeders per pen positioned side by side. Two feeders increased feed intake (P < 0·05) but had no significant effects on growth rate or feed conversion. With two feeders, the total time that feeders were occupied and the number of visits to feeders were both increased, while the number of enforced withdrawals from feeders and the amount of queuing were both decreased compared with the single feeder. In the second experiment, placing two feeders 2 m or more apart instead of side by side had no significant effects on feed intake or growth rate but improved feed conversion efficiency by 4%. This improvement was associated with fewer and longer visits to feeders but no difference in enforced withdrawals or queuing. In both experiments all treatments showed broadly similar patterns of diurnal feeding behaviour with two peaks of activity daily, a smaller peak pre-midday and a larger one in the afternoon. This pattern was influenced more by number of visits to feeders than by the mean duration of each visit. It is recommended that two single-space feeders are used in pens of 20 finishing pigs when non-pelleted diets are offered ad libitum and that feeders are sited some distance apart, not side by side.

Thirty-six 4-year-old Border Leicester × Romney (BR), Poll Dorset × BR (PBR) and Suffolk × BR (SBR) ewes balanced for litter size were assigned, 6 days after the mid-point of lambing (29 June 1989), to one of three sward surface height (SSH) treatments replicated twice (six ewes/1 ha paddock). The mean SSH during the 54 days of the trial were 2·6, 4·4 and 7·8 cm. Herbage organic matter intake (OMI) was measured using intraruminal chromium controlled release capsules during weeks 3, 4, 7 and 8 of lactation. Ewe OMI was not influenced by SSH during week 3 of lactation. During weeks 4 and 7 of lactation, ewe OMI increased with SSH up to a SSH of 7·8 cm. In week 8 of lactation there was no significant difference in OMI between ewes grazing the 4·4 or 7·8 cm SSH treatments, suggesting maximum intake at this stage of lactation is reached at 4·4 cm SSH. A maximum OMI of 2·6±0·1 and 3·0±0·1 kg OM/ewe per day (P < 0·05) was reached in week 4 of lactation on the 4·4and 7·8 cm SSH treatments, respectively. Ewes grazing the 2·6 cm swards had a relatively constant OMI throughout the experiment (range 1·8±0·1 to 2·0±0·1 kg OM/ewe per day). However, these ewes lost 10·5 kg of liveweight while those on the 4·4 and 7·8 cm SSH treatments maintained a constant liveweight. Sward surface height had no effect on ewe midside wool growth rate or mean fibre diameter (measured at day 60 of lactation) or on lamb liveweight gain. Ewe breed cross had no effect on ewe OMI or midside wool growth. Lambs reared by SBR ewes were heavier (20·4±0·5 kg) at day 60 of lactation than lambs reared by PBR (18·7±0·6 kg) or BR (18·4±0·6 kg) ewes (P < 0·05). These results suggest that for Winter(June)-lambing ewes in good condition a SSH of 4·0 cm under continuous stocking should ensure near to maximum ewe and lamb performance. However, ewes in good body condition at lambing stocked on a SSH down to 3 cm will rear lambs to acceptable weaning weights but at the expense of ewe liveweight loss and possible long-term carry-over effects on future productionif compensatory weight gains are not achieved during the spring pasture growth period.