Drought is probably the most important factor limiting crop yields worldwide, therefore it is not surprising that there has been continuing interest in the ways in which drought affects crop yield. Efforts have been concentrated in this area in the hope that it would prove possible to use a knowledge of drought physiology to provide a rational basis for the development of rapid methods of breeding drought tolerant cultivars, and also to help in the improvement of crop management for dry conditions. The last five years have seen some important reassessments of the underlying principles and concepts involved in plant response to drought and these will be outlined in this brief review. Some of these important shifts in emphasis have been highlighted by Kramer (1988), Passioura (1988), Schulze el al. (1988) and Boyer (1989), particularly in relation to the question of what measure of water stress is most relevant to plant function. As it is not possible to cover all aspects of drought physiology in a brief review of this nature, we highlight four topics where recent findings may have particular relevance to the improvement of drought tolerance in agricultural crops.

Maize was grown al the Regional Agricultural Research Station, Hathazari, Bangladesh during 1988/89 and 1989/90 to study fodder yield, cob growth and grain yield together with the contribution of pre-silking stem reserves to grain. Population densities were 33300, 44400 and 66600 plants/ha. Plants were either detopped after silking and pollen shedding, keeping 0, 2 or 3 leaves above the cob, or were left entire.

The results showed that the maize crop could successfully be detopped for fodder with little or no adverse effect on grain yield. Fodder yield increased with increased plant density and among the detopping treatments the highest fodder yield was obtained when the plants were detopped just above the cob. Cob growth followed a sigmoid pattern and the highest dry weight per cob was obtained from the lowest plant density and from entire plants. The number of cobs/m2 increased with increased plant density but detopping treatments did not give any significant difference in relation to densities. The number of grains/cob was highest with 33300 plants/ha but, among the detopping treatments, plants detopped just above the cob had the lowest number of grains/cob in both years. Weight of 1000-grain decreased with increasing plant density but it was increased by detopping plants just above the cob during 1988/89, although it was decreased in 1989/90. The highest apparent translocation of pre-silking reserves was obtained using densities of 44400 plants/ha but detopping treatments did not show any consistent effect, although the highest apparent translocation (20%) and harvest index (58%) were obtained from plants detopped just above the cob.

In experiments at Rothamsted in 1984–86, seven factors, each at two levels, were tested in factorial combination on winter barley (cv. Panda) grown as a third take-all susceptible crop. The factors were seed rate, a growth regulator prior to stem extension, amounts of N in spring, ‘winter’ nitrogen, an autumn insecticide, a fungicide applied to the seed (‘Baytan’) and a programme of fungicide sprays in spring and summer. Sowing 50% more seeds than normal increased the number of ears/unit area but had no effect on mean grain yield because grains were smaller. There were, however, significant, but unexplained, interactions between seed rate and the fungicide ‘Baytan’ applied to the seed. A growth regulator applied prior to stem extension had little effect on crop growth and no significant effect on grain yield. If sufficient N was applied in April there was little benefit from applying ‘winter’ N (30 kg/ha in November and again in February/March) except in 1985 when the amount of NO3-N in the soil, measured in the previous October, was lowest. Insecticide sprays applied in autumn to control the aphid vectors of barley yellow dwarf virus (BYDV) had no significant effect on grain yield but infectivity indices were below the threshold needed for treatment in each year. On average, ‘Baytan’ applied to the seed increased grain yield by 0·28 t/ha and this was associated with decreases in the severity of take-all. Over the three years, programmes of fungicide sprays, applied during spring and summer, increased grain yield by 0·92 t/ha but the mean response was largest where most N was applied.

The experiments also allowed the importance of interactions between different agronomic factors to be examined. A combined analysis of grain yields for all three years (based on 192 plot values) showed that only six 2- or 3-factor interactions, out of the 73 estimated, were significant (P < 0·05). Two of these interactions reflected variable responses to ‘winter’ N and fungicide sprays in the three years and three of the remaining four involved ‘Baytan’.

Additional plots of barley grown after oats had little take-all and yielded 1·14 t/ha more grain than similarly treated plots grown after barley. These responses were obtained despite evidence that oat residues had adverse effects on the growth of barley seedlings. Additional plots of barley grown after a bare fallow also had little take-all and gave even larger total yields (grain plus straw) than did barley after oats but the mean yield of grain was less than after oats because more of the dry matter after a fallow was straw. In 1984, when take-all was relatively slight, plots after a fallow gave even less grain than plots after barley (−0·77 t/ha) despite producing 2·12 t/ha more dry matter in grain plus straw.

Nitrate leaching from sheep-grazed pastures was measured over three winters between 1987 and 1990 at a lowland site in Wales. Losses from replicated plots of ryegrass/white clover (Lolium perenne/Trifoliwn repens) pasture that received no N fertilizer and from pure ryegrass pasture receiving 150–200 kg fertilizer-N/ha per year were determined using suction cup samplers. In Year 1, the average clover content of the grass/clover swards (April-September) was equivalent to 35% of total dry matter but declined during the course of the experiment and was only 4% in Year 3. Concentrations of nitrate and quantities of N leached from separate sampling positions exhibited a wide range of values and conformed to a log-normal distribution. In the first two winters, mean nitrate concentrations in soil water and quantities leached from grass/clover plots were greater than those from fertilized grass plots. In Year 3 the loss from the N-fertilized plots was higher and exceeded that from the grass/clover treatment. Total quantities of N leached during the three winters ranged from 6 to 33 kg N/ha from grass/clover plots and from 2 to 24 kg N/ha from fertilized grass plots. At no time did mean nitrate concentrations from either treatment exceed 11 mg N/litre. These values probably underestimate the true loss as the distribution of the main sets of water samplers did not provide adequate coverage of those areas of the plots where sheep congregated and which received high returns of excreta. These positions were characterized by particularly high leaching losses. Differences between losses from the two treatments and between years could only be partly explained by the decline in clover content and differences in stocking rates.

Stolon nodes of white clover were sampled monthly for 18 months from continuously grazed (set stocked) and rotationally grazed pastures in New Zealand. Both pastures were stocked at 22·5 ewes plus lambs/ha. Axillary buds were classified into viability categories using an incubation technique. On average, 54% of nodes had non-viable or dormant buds, 25% had axillary buds that were viable but non-active and 0·1% bore reproductive buds. Although 21% of buds emerged, only 8·5% of nodes bore live branches. Bud activity was strongly inhibited during the spring period of both years and this inhibition could not be removed by incubation.

The greatest bud viability occurred at node 4 (from the apex) although 30–40% of the buds at this position were non-viable. This pattern of variation changed somewhat with season. Some viable buds were present at nodes more than 20 from the apex but these were of low vigour. Secondary stolons had different patterns of bud activity from those of primary stolons until they comprised more than 16 nodes.

The results of 20 experiments which compared seed from July-planted crops with seed from earlier planted (April-May) crops either grown locally or from certified seed-producing areas are reported. The effects of dates of planting and harvesting of late-planted seed crops and seed storage temperature were examined in eight experiments. In five experiments, the effect of seed rate was studied on two weights of seed-tubers in the July-planted crops.

Tubers of several cultivars from the July-planted crops ended their dormancy close to the normal planting time of ware crops, 2–4 months later than tubers from spring-planted crops grown either locally or in areas certified for seed production. Tubers from seed crops planted in July produced fewer mainstems in ware crops than tubers from spring-planted crops, but effects on stem density were usually too small to have any significant effect on foliar ground cover and consequently on tuber yield. Date of planting the seed crops in July and date of harvest of the seed-tubers had no effect on sprout or ware crop growth in Record or Wilja. Ware tuber yields were similar for the different seed sources in 12 out of 15 experiments, suggesting little difference in the performance of tubers from seed crops planted early and late in the season.

At equivalent seed rates, small seed-tubers (12·5 ± 2·5 g) from July-planted seed crops usually produced more stems and tubers and greater ware yields than larger tubers (37·5 ± 2·5 g). For the smaller seed, the results suggested optimum seed rates for ware tuber yields of c. 0·88, 0·88 and 0·59 t/ha for Wilja, Record and Saturna, respectively. These are substantially lower than those currently used in the UK. For the larger seed, rates of c. 2·64, 2·64 and 1·76 t/ha were adequate for Wilja, Record and Saturna, respectively.

Twenty genotypes of chickpea, selected according to their performance in trials conducted during 1988/89 in saline fields having equal concentrations of SO4 and Cl, were studied during 1989/90 at Karnal, India. The Cl-dominated soils were divided into microplots with ECe values of 2, 4, 6 and 8 dS/m. Germination, although delayed at ECe 6 and 8 dS/m, was not inhibited up to ECe 8 dS/m. The sensitivity of all genotypes increased with plant growth and greater salinity. Plant growth was most adversely affected by a salinity of ECe 8 dSm. 100-seed weight was less affected by salinity than other yield components, namely, number of pods/plant, number of seeds/plant and seed yield/plant. Chlorides had a more severe toxic effect when present in domination than when they were balanced by an equal concentration of sulphates. Genotypes varied in their tolerance of Cl-dominated salinity. Cultivars ICCC32 and 1CCL86446 showed the most tolerance to Cl-salinity, having > 0·75 Mean Tolerance Index values for all eight characters, followed by ICC5003, ICC10575, ICC12908 and ICC12926. ICC12928 had the highest mean tolerance index value for number of pods/plant and highest threshold value and ICC4953 the highest mean tolerance index for 100-seed weight, indicating their better tolerance at flowering and maturity respectively. Therefore, crossing genotypes of the latter with the former group should produce some recombinants with improved tolerance to salinity at all growth stages and at maturity.

Forty-one experiments were carried out in England and Wales between 1983 and 1985 to compare ammonium nitrate and urea as N top dressings for multicut silage.

The results showed that relative dry matter yields from the two fertilizers differed considerably between sites. However, compared with ammonium nitrate, there was a mean yield decrease with urea of 2% at the first cut and 5% at the second cut. Mean herbage N contents and apparent N recoveries were lower with urea than with ammonium nitrate at both the first and second cuts. It appeared, at least for first-cut dressings applied in early spring, that urea effectiveness increased with the amount of rain falling within 3 days of fertilizer application. The effect was much less obvious at the second cut.

Urea effectiveness was not markedly reduced on soils of high pH or light texture, where higher ammonia volatilization losses might have been expected.

Treatments which lowered the rate of ammonia volatilization from surface-applied cattle slurry were evaluated in three field experiments during 1989/90 at the Agricultural Research Institute of Northern Ireland. The relative effects of separation, dilution with water before application and washing with a water spray after application, on ammonia volatilization rates were compared over 4 days using ventilated enclosures. As the loss rate of ammonia during the first 4 days after application was highly correlated with the loss rate over the first 4 hours after application, the loss rate in the first 4 hours was used as a rapid method to assess the efficacy of various combinations of separation, dilution and acidification. Ammonia volatilization loss rates from treatment combinations were compared to the loss rate from whole slurry.

A 50% decrease in ammonia volatilization was achieved by separation through a 0·4 mm mesh, separation through a 10·0 mm mesh plus dilution with 86% by volume of water, or separation through a 2·0 mm mesh plus washing with 53% by volume of water.

A 75% decrease in ammonia volatilization was achieved by acidification to pH 6·5, or combinations of separation and dilution, namely separation through a 0·4 mm mesh plus 50% dilution or separation through a 5·0 mm mesh plus 100% dilution.

A 90% decrease in ammonia volatilization was achieved by acidification to pH 6·0, dilution by 50% plus acidification to pH 6·5, or separation through a 0·4 mm mesh plus acidification to pH 6·5.

The effects of separation and acidification on ammonia volatilization were cumulative. There was a highly significant inverse linear relationship between ammonia volatilization rate and volume of nitric acid used. Adding 10 M nitric acid at 1·4% by volume lowered volatilization by 75% of that from whole slurry and increased the nitrogen content of the slurry by 2 g N/l, making it a more balanced NPK fertilizer for cut swards. Because the variable value of cattle slurry is due to loss of nitrogen as ammonia and smothering of the sward with slurry solids, a combination of separation or dilution to lower sward contamination and acidification with nitric acid to lower volatilization may be the best practical option for improving cattle slurry as a fertilizer for cut grass.

Combinations of treatments of cattle slurry, which have previously been shown to have lowered ammonia volatilization by 75% of that from whole slurry, were compared in a field experiment with perennial ryegrass at the Agricultural Research Institute, Hillsborough in 1990. Whole slurry and slurries obtained by separation through screens with mesh sizes of 5·0, 3·0, 1·1 and 0·4mm were acidified with 1·4% by volume of 10M nitric acid. Slurries separated through 5·0 and 0·4 mm meshes were diluted 100% and 50% respectively by volume with water and amended with calcium nitrate to supply the same amount of nitrogen as the nitric acid. All slurry combinations were surface-applied at 100 kg (NH4+-N + NO3−-N)/ha to different plots on three occasions during the growing season. Volatilization of ammonia was measured on adjacent plots using ventilated enclosures. By comparison with inorganic fertilizer treatments, assessments were made of the efficiency of slurry nitrogen for herbage production, the fertilizer value of phosphorus, potassium and sulphur in the slurry, and the effects of slurry on silage quality.

The mean ammonia volatilization over all applications of acidified slurries was 17% of that from whole slurry. The two separation plus dilution treatments, however, only lowered volatilization to about half of that from whole slurry. Within separation plus acidification treatments, the trend was for dry matter yield and nitrogen offtake to increase as mesh size decreased down to 1.1 mm. For slurries separated through 5·0 and 0·4 mm meshes, the effect of the additional dilution treatment on dry matter yield and nitrogen offtake was similar to the effect of acidification. Efficiency of nitrogen offtake over all three cuts, relative to inorganic fertilizer nitrogen, was 54% for acidified whole slurry and 88% for slurry separated through a 1·1 mm mesh and acidified. Denitrification in nitric-acid amended slurries and sward contamination from slurry solids may have been responsible for lowering nitrogen efficiency.

Chemical analyses of herbage showed that the concentrations of phosphorus, potassium and sulphur from the slurry treatments were comparable to those in herbage receiving the recommended rates of inorganic fertilizers and were sufficient for optimum yield. Slurry application had a detrimental effect on the fermentation quality of the ensiled herbage. Chemical analyses indicated that fermentations from two slurry treatments at the first cut and all slurry treatments at the third cut were poorer than those on treatments receiving the equivalent rate of inorganic fertilizer nitrogen.

Twelve yearling Droughtmaster heifers were grazed on a tropical grass/legume pasture in north-east Australia for 1 year in four paddocks, each of 4 ha. One paddock was fertilized annually with 10 kg P/ha (treatment HPF). The other three paddocks had low soil P and low pasture P status. In one of the low soil P paddocks, P was added to the drinking water (treatment HPS) but no supplement was provided in the other two paddocks (treatments LP1 and LP2). In February, May and August 1989, the P kinetics of the heifers were monitored following oral administration of ruminal chromic oxide capsules and i/v injection of 32P. At the same times, oesophageal-fistulated steers were grazed in the paddocks to obtain extrusa samples to estimate the P content and digestibility of selected forage.

The liveweight gain per heifer was higher in the HP treatments (196 kg) than in the LP treatments (152 kg). The intake, absorption and excretion of P were closely related but the faecal excretion of P relative to P intake was less in heifers with than without P supplement. The P intakes of the LP heifers were about half the quantities recommended in the nutrient feeding standards for the observed food intakes and weight gains. The coefficients of P absorption in the LP heifers in May and August were high (> 0·8) and did not decrease with age or liveweight. The coefficients of P absorption of heifers given supplementary P were as high as those for LP heifers, but those of heifers on the fertilized pasture were significantly lower in May and August. The endogenous faecal excretion of P was related to P intake but more closely to dry matter intake and plasma inorganic P together. The reasons for the difference between observed P intakes of LP heifers and recommended requirements are discussed in relation to the assessment of P requirements of grazing cattle. The results from this experiment suggest that recommended P requirements are higher than necessary to sustain the measured liveweight gains of cattle on grass/legume pastures in the tropics.

Four Friesian bulls with ruminal and duodenal cannulae were used in a 4 × 4 Latin square experiment to study the effects of lactic acid (LA) on rumen fermentation and microbial protein synthesis. On a dry matter (DM) basis (g/kg), the basal diet comprised grass silage (700), barley (240) and rapeseed meal (60) and it was given at the rate of 7·1 kg DM/day. LA was infused continuously into the rumen at the rates of 0 (L0), 40 (L40), 80 (L80) or 120 (L120) g/kg basal diet DM.

The molar proportion of propionate in the rumen volatile fatty acids (VFA) increased linearly (P < 0.001) and that of acetate, isovalerate, caproate (P < 0.01) and isobutyrate (P < 005) decreased linearly with an increasing rate of LA infusion. At the same time there was a linear decrease (P < 0.05) in the number of rumen protozoa. When the metabolic fate of infused LA was calculated on a molar basis, 0.21 of lactic acid was converted to acetate, 0·52 to propionate and 0.27 to butyrate.

Infusion of LA into the rumen had no effect on the site or extent of the digestion of basal diet organic matter (OM) and neutral detergent fibre (NDF). LA diets tended to have a lower microbial N flow at the duodenum (71·4 v. 85·8 g N/day) and lower synthetic efficiency in the rumen (14·4 v. 20.4 g N/kg OM apparently fermented) when compared with the control diet. The ratio of duodenal non-ammonia N to N intake was highest with the control diet and lowest with L40, the effect of the LA rate being quadratic (P < 0·05). The results suggest that propionate was the main end-product of lactic acid fermentation in the rumen with the grass silage based diet. Lactic acid had no value as an energy source for microbial protein synthesis.

Dorper sheep are raised in extreme desert areas. Body mass loss and body fluid shifts were measured in Dorper rams denied water for 4 days and offered only wheat straw. The rams lost 16·3% body mass, 22·0% total body water volume, 35·1 % extracellular fluid volume and 41·7% plasma volume. On first drinking following dehydration, Dorpers were able to consume 19·7% of their dehydrated body mass and 100·3 % of their body mass loss. It was concluded that Dorpers can survive in harsh deserts through their ability to withstand dehydration and quickly replenish body mass losses when water becomes available.

Aspergillus oryzae fermentation extract (AO; 2 g/day) was added to the diet of sheep fed grass hay supplemented with 30 or 70% barley. AO decreased the proportion of propionate in rumen volatile fatty acids with both levels of supplementation (146 and 163 mmol/mol with AO v. 157 and 186 mmol/mol with no addition for the low and high barley diets respectively; P < 0·05), and also caused a small reduction in L-lactate concentration in the high barley diet (1·15 mM v. 1·43 mM in the absence of AO). Rumen pH was not changed significantly. Total viable bacteria in the rumen were stimulated with AO (2·12 and 2·46 v. 0·97 and 1·80 × 109/ml respectively). Numbers of cellulolytic bacteria and ciliate protozoa were unchanged. Hay suspended in nylon bags in the rumen tended to be degraded more rapidly with AO, but the effect was not statistically significant. Neither barley nor AO significantly altered the plasma concentrations of glucose, urea, insulin, gastrin or cholecystokinin. It was concluded that the effects of AO were generally similar for both levels of barley supplementation and that changing the composition of the diet had no detectable effect on the measured indicators of nutritional status in blood.